Docstoc

Equity Valuation for Analysts and Investors

Document Sample
Equity Valuation for Analysts and Investors Powered By Docstoc
					 Equity
Valuation
for      Analysts
&       Investors
A Unique Stock Valuation Tool
for Financial Statement Analysis
      and Model-Building



                Jim Kelleher




New York Chicago San Francisco Lisbon London Madrid Mexico City
   Milan New Delhi San Juan Seoul Singapore Sydney Toronto
Copyright © 2010 by The McGraw-Hill Companies, Inc. All rights reserved. Except as permitted under the
United States Copyright Act of 1976, no part of this publication may be reproduced or distributed in any form or
by any means, or stored in a database or retrieval system, without the prior written permission of the publisher.

ISBN: 978-0-07-175952-6

MHID: 0-07-175952-2

The material in this eBook also appears in the print version of this title: ISBN: 978-0-07-163923-1,
MHID: 0-07-163923-3.

All trademarks are trademarks of their respective owners. Rather than put a trademark symbol after every
occurrence of a trademarked name, we use names in an editorial fashion only, and to the benefit of the trademark
owner, with no intention of infringement of the trademark. Where such designations appear in this book, they
have been printed with initial caps.

McGraw-Hill eBooks are available at special quantity discounts to use as premiums and sales promotions, or
for use in corporate training programs. To contact a representative please e-mail us at bulksales@mcgraw-hill.
com.

TERMS OF USE

This is a copyrighted work and The McGraw-Hill Companies, Inc. (“McGrawHill”) and its licensors reserve all
rights in and to the work. Use of this work is subject to these terms. Except as permitted under the Copyright Act
of 1976 and the right to store and retrieve one copy of the work, you may not decompile, disassemble, reverse
engineer, reproduce, modify, create derivative works based upon, transmit, distribute, disseminate, sell, publish
or sublicense the work or any part of it without McGraw-Hill’s prior consent. You may use the work for your
own noncommercial and personal use; any other use of the work is strictly prohibited. Your right to use the work
may be terminated if you fail to comply with these terms.

THE WORK IS PROVIDED “AS IS.” McGRAW-HILL AND ITS LICENSORS MAKE NO GUARANTEES
OR WARRANTIES AS TO THE ACCURACY, ADEQUACY OR COMPLETENESS OF OR RESULTS
TO BE OBTAINED FROM USING THE WORK, INCLUDING ANY INFORMATION THAT CAN BE
ACCESSED THROUGH THE WORK VIA HYPERLINK OR OTHERWISE, AND EXPRESSLY
DISCLAIM ANY WARRANTY, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO IMPLIED
WARRANTIES OF MERCHANTABILITYOR FITNESS FORAPARTICULAR PURPOSE. McGraw-Hill and its
licensors do not warrant or guarantee that the functions contained in the work will meet your requirements or
that its operation will be uninterrupted or error free. Neither McGraw-Hill nor its licensors shall be liable to
you or anyone else for any inaccuracy, error or omission, regardless of cause, in the work or for any damages
resulting therefrom. McGraw-Hill has no responsibility for the content of any information accessed through
the work. Under no circumstances shall McGraw-Hill and/or its licensors be liable for any indirect, incidental,
special, punitive, consequential or similar damages that result from the use of or inability to use the work, even
if any of them has been advised of the possibility of such damages. This limitation of liability shall apply to any
claim or cause whatsoever whether such claim or cause arises in contract, tort or otherwise.
   Marie. . .


. . . and Angus,
Jack, & Wallis
This page intentionally left blank
          CONTENTS




Acknowledgments     vii

Introduction   xi


   PART 1 INCOME STATEMENT PRESENTATION 1
 Chapter 1 Phase 1: Income Statement and Margin Model, Part 1      5
 Chapter 2 Phase 1: Income Statement and Margin Model, Part 2     39
 Chapter 3 Phase 2: Segment Modeling of Revenues                  63
 Chapter 4 Phase 3: Segment Operating Income and Percentage-
           of-Difference Modeling                                  73
 Chapter 5 Phase 4: The Workbench, Part 1                          83
 Chapter 6 Phase 4: The Workbench, Part 2                         103
 Chapter 7 Ordinary Least Squares Regressions and
           Normalized Earnings                                    121


   PART 2 RATIO AND VALUATION WORKSHEET                     137
 Chapter 8 Ratio Analysis, Part 1: Internal Liquidity and
           Operating Efficiency                                    143


                                                                    v
vi   •   Contents



 Chapter 9 Ratio Analysis, Part 2: Return Ratios and Cash
           Flow Ratios                                                  163
Chapter 10 Historical Comparable Valuation                              187


     PART 3 STOCK VALUE WORKSHEET             219
Chapter 11 Present Value Modeling and the Stock Value Worksheet         223
Chapter 12 Discounted Free Cash Flow: Setting the Table                 235
Chapter 13 Discounted Free Cash Flows: Two Methods                      251


     PART 4 RELATIONAL VALUATION: THE INDUSTRY
            MATRIX WORKBOOK AND PEER DERIVED VALUE                277
Chapter 14 Price and Performance Analysis                               283
Chapter 15 Simple Average and Market-Weighted Comparisons               303
Chapter 16 Peer Derived Value                                           331

Conclusion: Dollar Value of the Asset   359

Bibliography 363

Index 365
          ACKNOWLEDGMENTS




I have spent most of my financial career and all of my analysis career at Argus
Research, an independent equity research firm founded by Harold Dorsey in
1934. My nearly 20 years at Argus have been marked by unprecedented tumult in
the financial world, which the firm has navigated steadily and professionally. I
am indebted to Argus and the Dorsey family for creating a culture that invites
and rewards independence, self-motivation, and innovation while maintaining
consistently high standards in investment analysis and portfolio management.
      Argus CEO and president John Eade hired me at Argus nearly two decades
ago. His support for my succeeding him as Director of Research strengthened my
confidence in my ability to write this book. Richard Cuneo, Director of Opera-
tions at Argus, is another mainstay at the company and in my professional devel-
opment. Sharon Dorsey Wagoner and Fern Dorsey serve ably as heads of Argus
Investors’ Counsel and Vickers Stock Research, respectively.
      The composer, musician, and private wealth manager Dana Richardson
worked as an analyst at Argus earlier in the decade. Discussions with Dana were
the seedbed for the set of concepts that evolved into Peer Derived Value. I am
indebted to Wendy Abramowitz, skilled analyst and top stock-timer, for intro-
ducing me to comparable historical analysis as well as key themes in technology-
sector analysis. Jim Solloway, CFA, former Chief Economist and Director of

                                                                             vii
viii   •   Acknowledgments



Research at Argus, taught me the fundamentals of OLS regressions for smoothing
long-term growth rates. Bob Becker, CFA, was my partner in innumerable proj-
ects at Argus, and his counsel was always excellent—particularly his guidance on
passing the CFA (“take the practice tests ‘till you’re climbing the walls”).
      Other analysts whose work directly or indirectly influenced the body of
knowledge underpinning this work include Chris Graja, Joe Bonner, Kevin Cala-
brese, Suzanne Betts, Bill Selesky, Erin Smith, David Toung, Martha Frietag,
David Kerans, Phil Weiss, John Staszak, David Ritter, and Gary Hovis, dean of
the electric utility analyst community. Kevin Tynan is the only automotive ana-
lyst in captivity who can break down a carmaker’s pension obligations while
simultaneously rebuilding a carburetor; he has been my chief guide in decoding
the mysteries of Excel.
      The person identified with the quote that opens the book is Betty “B.J.”
Edwards, long-time Chief Editor at Merrill Lynch. Seemingly with a few pencil
strokes, she helped transform my compositional skills from a scattering of con-
cepts to a well-ordered file cabinet of handy rules and strategies. Though we’ve
never met, I am indebted to Frederick Crews, author of The Random House Hand-
book, who showed that a light touch, far from degrading the seriousness of a
work, helps make valuable lessons indelible (“. . . who but that orderly’s mother . . .”
indeed).
      Sophie Efthimiatou was the McGraw-Hill editor who, on the strength of a
recommendation and my pleas over a cup of coffee, became my early champion.
She helped me tighten and hone my proposal until it became the book I’d spent
20 years preparing to write. Jennifer Ashkenazy, CPA, gave the accountant’s
thumbs-up to the project. Morgan Ertel was indispensible in suggesting the
restructuring of the book into its four-part structure and advancing the project
from rough manuscript to finished product. Daina Penikas guided me through
the copyediting and proofreading process with great good humor.
      At my nephew James’ wedding in August 2008, I ran into my cousin Trish
Pignataro, CPA, who immediately started giving me the needle: “When are we
going to see that book?” and “I always thought you were going to write a book.”
At a time when I was mulling just such a project, this goad proved to be a tipping-
point event. On her behalf, I wrote a book that only an accountant could love.
      My wife Marie has nudged me out of one rut after another over the years
and is largely responsible for my semi-respectable state. While I was writing at all
odd hours, the kids managed to keep the mayhem at sub-Bedlam levels. The year
this book was written, 2009, was without doubt the busiest of my corporate life.
Despite all the time stolen from family by book and work, no one complained
and everyone was supportive.
                                                          Acknowledgments   •   ix



      My delusions of mathematical grandeur notwithstanding, I might burn out
the batteries on my calculator trying to figure out what I owe Rich Yamarone,
Bloomberg economist and author of the invaluable The Traders’ Guide to Key
Economic Indicators. Most directly, he introduced me to his editor, with a kind
word and a recommendation, and brought to bear his considerable reputation in
the financial publishing world on my behalf. Without his nod, there would likely
have been no book. Less directly and more importantly, he demonstrated that the
guy in the next office could have the audacity and tenacity to conceive, design,
write, and publish a highly useful text for the financial services industry. I am
eternally indebted to Rich, unless he actually asks me for money.
This page intentionally left blank
           INTRODUCTION




T
“ here are many ways to skin a cat.” So spoke a senior editor at Merrill Lynch,
explaining her unerring ability to wring clarity and concision from the varied but
consistently gnarled prose styles of the equity analysts. Bad writing in all its iter-
ations—stuffy, unreferred, thin, florid, lacking in segue or sequence—never
boxed her into a corner; she knew multiple ways out.
      Despite the shortcomings in their writing styles, the analysts with which we
worked too seemed to skin cats—or analyze stocks—in a lot of different ways.
However varied their approaches, they always sought to derive the same thing:
the dollar value of the asset. Editing and analysis, I was learning, traveled many
paths to a single outcome.
      Having barely digested this wisdom, I left for Dallas where I wrote and
typeset an automotive trade journal. Back in New York in the early 1990s, I
returned to financial editing while being charged with a fair amount of writing.
Surmising that editing would never pay New York City rents, I became a financial
analyst early in the decade. Eventually I entered the CFA (Chartered Financial
Analyst) program and had my financial analyst charter in hand by 1999.
      As a former English major, I was tabula rasa for the business world and had
no bad habits to unlearn. And armed with the excellent knowledge garnered in
the three-year CFA program, I felt prepared for the measured and precise long-
term valuation of assets.
                                                                                    xi
xii   •   Introduction



      The first thing I learned, however, was that those assets wouldn’t stand still,
which—as far as investors were concerned—meant that long-term valuation be
damned. The movement in equities I observed rarely correlated to longer-term
trends within their peer groups or even within their own financial histories.
Instead, stocks appeared to be dancing to their own tune, and they took a step
this way or that each time a new product was introduced (or flopped), the com-
petitive landscape underwent secular or cyclical changes, regional market surged
or retreated, and so on.
      The longer-term movements in stocks tended to supersede these daily gyra-
tions. Yet some companies meaningfully diverged from trends when their com-
petitive position cumulatively changed, or their asset portfolio was overhauled,
or new entrants ate their lunch. Analysis, I was learning, involved blending cycli-
cal, secular, and company-structural events into the mix without spoiling the
soup. Within the market noise, I was eventually able to discern the signal: con-
sistent profits. Just as reliable earnings differentiated the successful companies
from the pretenders, reliably modeled income flows and cash streams informing
the valuation process became the best guide as to whether a company was suc-
cessfully navigating industry transitions or succumbing to competitive
pressures.
      In an earlier, seemingly more staid time, swaths of companies devoted to a
single industry—paper, say, or railroading—could be valued from the top down,
with a focus on return on equity (ROE) and reasonably consistent variations
from growth in gross national product (GNP). But as competition grew more
intense and global, as companies—even those competing around a single com-
modity—increasingly pursued their own paths to profits, conclusions derived
from top-down analysis veered further and further from market reality. As top-
down analysis fell from favor, bottom-up analysis not only proliferated, it pro-
duced its own mantra: go granular. It was no longer enough to produce rounded
earnings per share (EPS) estimates based on general and long-term trends. EPS
forecasts and other inputs, such as cash flow, needed to reflect the myriad forces
and fast-flowing information driving line items on a segment-by-segment and
even subsegment basis.


Finding and Refining the Approach
How to derive those income flows and cash streams? As I immersed myself in
analysis and got to know my colleagues, I realized there was no single template
for calculating income and cash flow. Nor was there a single reliable method for
                                                                Introduction   •   xiii



valuing equities. This went well beyond inherent differences from industry to
industry; within single industries and even within narrow niches, income model-
ing and asset valuation approaches varied widely. Everyone seemed to be doing it
their own way. Some analysts had inherited models or relied on the advice of
mentors. Other analysts, dissatisfied with inherited wisdom, were fashioning
their own models, particularly for companies and industries that were previously
nonexistent but now seemed to be at the center of the market’s obsession.
      Almost without realizing I was doing so, I felt my way toward a sustainable
analysis model. Skittering at the edge of consciousness were a few constants. In
the market there are no rights and wrongs, no light under the bushel basket; the
market is driven by perceptions and realities, and a comprehensive valuation
scheme must accommodate both. If earnings and cash flow drive the valuation
process, they must be precisely modeled and then seamlessly integrated into valu-
ation analysis. The valuation model must accommodate both minute-to-minute
developments and long-term trends. And no pertinent data point could be
orphaned, marginalized, or left behind.
      As this overriding objective began to coalesce, I realized the scheme was
easier envisioned than executed. The variety of tools for modeling and valuation,
though individually useful, are in their abundance the analysts’ greatest chal-
lenge. Certainly, the industry tool kit was plentiful. From the CFA process, related
courses, mentors, and colleagues, I learned financial statement modeling, com-
parable historical valuation, discounted free cash flow valuation, industry analy-
sis, and a host of other neat tricks. What I needed, though, was a way to organize
all these inputs into a single stream that all contributed to the final output.
      How to rank and prioritize among them? For the modeling of income and
cash flows, there was no single template; ditto for the determination and applica-
tion of growth rates. Historical comparables analysis captured the past valuation
experience with precision but carried within itself the warning that it could not
look too far ahead.
      The various present value schemes, such as the dividend discount model
and discounted free cash flow (DFCF) valuation, were ingeniously constructed
to capture the long-term value of the asset. But what about the here and now? If
DFCF signaled that a reliable growth rate for the company was 6 percent, but then
that company indicated a glitch in its production—a fire at a motors plant in
Jakarta, say—what next? Shave the growth rate to 5.875 percent? And for how
long?
      Valuation within the peer group presented even bigger challenges. Seem-
ingly nothing could be more vital or telling than peer valuation, yet it had a
surprisingly touchy-feely aspect. What’s more, peer evaluation’s implicit message
xiv   •   Introduction



(e.g., “This stock used to trade at a premium, it is now at a discount, so do some-
thing”) really carried no guidance on how to proceed.
       As time passed, knowledge accrued. When one is seeking to prioritize
among schemes, when one is looking to assign each value its own gradient on the
valuation curve, nothing substitutes for experience. More specifically, there is no
better way to learn the art of valuation than enduring the humbling experience
of forecasting great things for a stock only to see it sink (or watching the over-
looked asset soar out of sight). You begin to calibrate, a term defined as adjust-
ments based on recent experience but that in practice entails using past failures
to steer you closer to the vital truth. Gradually ego subsides; you stop fighting the
market and begin to work with it. You fit your scheme to encompass all of the
market’s information.
       The analyst’s challenge then begins to compound into a linked series of
procedures. We commence by estimating income and cash flows reliably. We cast
these estimated values into the web of historical inputs and value relationships.
We incorporate industry data where appropriate. And we weave this information
consistently into the valuation process without leaving any loose ends. As much
as possible, we seek to systemize the market’s valuation processes and then rank
and weight them. Even while establishing and enacting this dry and clinical pro-
cess, the analyst must incorporate the market’s chaos and dynamism, wherein
hunches and rumors can sometimes supersede rigorous valuation process. As
various goals and themes intermingle, the challenge becomes the practical and
consistent application and interaction of the various information inputs needed
to arrive at a value forecast.
       The developing analyst is immersed in and eventually becomes conversant
in the various theoretical approaches to asset valuation. In the end, the analyst
serves masters—research directors, portfolio managers, and ultimately the end
user or asset owner—far removed from financial academia. The phone rings;
steps are skipped; compromises are made. The analyst simply needs to value that
asset; few are interested in his or her process. Our task is not to argue financial
theory but to deploy it. So, we won’t, for example, defend or seek to upend such
widely accepted industry verities as capital asset pricing model (CAPM); we won’t
even explain it much. We’re going to take it as a given and simply put it to
work.
       Gradually, you arrive at the realization that estimating the dollar value of
the asset is not so much valuation theory application as it is valuation choreog-
raphy. The model needs to be supple and responsive enough that, if an input
changes, an entire chorus line of data points kicks in time. In a real-world exam-
ple, if an analyst changes an assumption about current-quarter pricing for
                                                                  Introduction   •   xv



second-generation mobile handsets at Motorola, the information needs to ripple
across the current and next-year income statement, up through comparable and
discounted free cash flow valuations, and into the calculated dollar-based fair
value.


An Overview of Equity Valuation for Analysts and Investors
On the one hand we’ve acknowledged that the analyst can follow many paths to
deriving dollar value of the asset; on the other, we’ve constructed a fairly specific
approach to building the model. Now, how do we reconcile the two? We won’t
take a “my way or the highway” approach; but working within our system pro-
vides a functioning and (double emphasis) beginning framework for financial
statement modeling and valuation analysis. Our goal is to provide a basket of
unified concepts so the self-directed analyst can construct his or her own model.
We’ll show you exactly how we do it while leaving room for variation as the
maturing analyst spreads his or her wings.
      Learning and teaching, while sharing some points in common, are very
different processes. If our goal is to have you learn how to apply modeling and
valuation technique, we’ll need to teach you specifically how to apply this in the
format of an excel workbook. In writing this book, I was struck with the chal-
lenge facing any instructor standing before a group of students, each with varying
degrees of intelligence, experience, and willingness to learn.
      Like that instructor, we’ll begin with a lot of hand-holding and the assump-
tion that even the most rudimentary formula and application must be thoroughly
explained. As such, the information and instructions that are offered in this book
will be accompanied by a level of exacting detail. Most of these exhaustive expla-
nations can be found in the very first section of this book, which details how to
build the modeled P&L. As we proceed, we’ll assume everyone is learning at the
same rate, and this almost paint-by-number level of detail will recede. We will
further assume the modeler has developed some familiarity with the workbook
and worksheet, and our instructions, while never cursory, will become less
detailed. As the book moves along, we’ll gradually reduce the accompanying
detail around every Excel formula.
      The book is concerned with two main themes: modeling and valuing. It is
structured in four parts: financial statement modeling; comparable historical
valuation; discounted free cash flow valuation; and relational valuation. Each of
the four parts begins with an opening essay, followed by multiple chapters. The
biggest section, financial statement modeling, has seven chapters; the three other
xvi   •   Introduction



parts have three chapters each. The text is book-ended by this introductory chap-
ter as well as a concluding chapter.
       As in any interconnected whole, concepts in any one chapter might argu-
ably be better suited for inclusion in another; but that would risk ripping the
fabric created by other relationships. In every chapter we begin with a discussion
of the topic, including the changing currents and fast-formulating priorities that
are shaping each topic as we speak.
       In a first introductory chapter, we describe the real-world challenges inher-
ent in modeling and valuation, describe the basic structure of the book, and
discuss our processes.
       Financial statement modeling, and most particularly income statement
modeling, is the topic of the book’s first section, encompassing seven chapters. A
first step in the valuation process is to build an income statement—forecast from
five to eight quarters out—that can incorporate company developments, industry
trends, and our best estimate of what the future will hold based on past practice
and experience. Anyone in the field has encountered many of volumes on valua-
tion. Income statement modeling gets second-class status on the premise that it’s
all percentage-of-revenue compilation. In fact, investors cannot reliably value the
asset if the financial statement model is not nuanced and comprehensive and
provides all the information possible.
       In practice, what we call the income statement presentation encompasses
the income statement model, but it also covers the accompanying margins, ratios,
segment data, and industry detail that enable more precise modeling. One feature
of this book is the recognition that the mundane and atypical can distract us
from the core task of valuation. Hence, in subsequent chapters we spend some
time on the exceptions—modeling foreign companies, accommodating stock
splits, and so on—that can disrupt the valuation process.
       We wrap up this long first section by demonstrating means of calculating
smoothed growth rates and normalized earnings—tools to better assess perfor-
mance across the various points in the economic cycle. A key danger in the valu-
ation process is the inability to reliably adjust for the economic cycle. Unpredictable
as it is, the economic cycle at one stage or another is continually impacting com-
panies. The last chapter in Part 1 provides some tools to accommodate these
cyclical forces.
       After a rigorous discussion of the modeling process, we move onto a com-
prehensive discussion of common—and proprietary—tools for equity valuation.
In Part 2, we discuss comparables historical valuation—that is, the use of histori-
cal price-relationship data and modeled inputs to derive asset value. The histori-
cal comparables chapters also include various useful ratios, some of which figure
                                                              Introduction   •   xvii



directly in individual asset value decision, some of which inform the industry
valuation framework, and some of which subjectively influence the valuation
decision.
       Part 3 is devoted to our take on present value modeling, specifically the
“discount to the firm” flavor of discounted free cash flow valuation. We examine
the risks inherent in this method, specifically DFCF’s implicit reliance on return
on equity at a time when accounting regulations and corner-office practice are
degrading the very validity of stockholders’ equity. We also use this format to
first discuss incorporating various inputs into determining value of the assets on
a risk-adjusted basis.
       In Part 4, we use the individual equity workbooks we’ve created to build
and populate an industry matrix. In it we can track industry data and perfor-
mance of the equity and its peer group on a simple average and weighted basis
and construct various alerts to capture gains or limit losses. The industry matrix
also provides a fulcrum for beginning analysis of the asset within its industry
group, along with techniques for market-weighting returns.
       Concluding Part 4 of the book, we address what we deem to be an industry
shortfall by explaining our method for peer-group relative value, called Peer
Derived Value.
       In the conclusion we briefly discuss the role of modeling and valuation
analysis within the analyst’s role.


How to Best Use This Book
How good is a newborn model? About as useful in the workspace as a newly
minted college graduate—which is to say that it is more likely to knock over the
coffee on your desk than it is to increase sales. College may not bestow a lot of
practical information to young people, but it does teach them how to learn how
to learn (at least we hope it does). Similarly, our wet-behind-the-ears model is
well-intentioned but awkward, not to mention alarmingly deficient on day one
in real-world common sense. But it is structured to accommodate ever more
information. A few months into the job, our recent college graduate may surprise
us with fresh insights and new energy brought to a familiar task. Similarly, our
model is designed to incorporate new inputs in the formulation of investment
opinion and determination of investment value.
      The new model must be structured to gather more data, so along the way
we’ll elaborate steps to enable an ever-more-granular approach. It also must cali-
brate and, finally, replicate. By that I do not mean that the new model just needs
xviii   •   Introduction



to be able to spawn like models for like (or even unlike) companies. It also must
be able to add a new measurement and valuation period (typically one year,
divided into quarters) without needless repetition of steps. It must be able to
reflect changes in company data presentation, something that happens a lot more
often than you’d think. Remember, our model lives in a world of ringing phones,
urgent e-mails, tense morning squawk sessions—no ivory tower, all business.
      One of the biggest challenges for analysts working in the real world is bal-
ancing the rigorous application of theory and process with shortcuts. When mod-
eling a balance sheet, the analyst can model each account in accord with the line
items in the cash flow statement—or he or she can increase all balance sheet
accounts uniformly at forecast gross domestic product or the asset’s historical
growth rate. Whenever possible, we describe the more rigorous process as well as
the shortcut. Again, our goal is not to bog you down in process or theory but to
help you build the model. Sometimes, the choice of formal process versus the
shortcut is related to your position in the value chain. The buy-side analyst
charged with keeping an eye on entire industries and sectors may make different
choices and compromises than a sell-side analyst charged with monitoring a tidy
group of 12 or fewer stocks.
      Modelers need to respond to new real-time information inputs; the model
is built to accommodate company information as it is issued. For instance, a
company typically may report its results 20 days after quarter end and use a
somewhat amended or modified income statement. Sometime later, within a
45-day window, it will issue its formal quarterly financial results within the 10-Q
format, and this income statement may be more detailed and nuanced. But if you
wait for this later input, you’ll be lagging a market that has already digested and
moved on from the real-time information issued on day 20.
      Once you’ve completed a full company modeling, don’t admire it too much:
the company has a fair chance of changing its reporting style. This may reflect
maturation of a one-time growth company, appointment of a new chief financial
officer’s competitive concerns, or a response to changes mandated by the U.S.
Securities and Exchange Commission. As much as possible, we build the overall
valuation model to seamlessly accommodate such changes, which typically occur
within the Income Statement Presentation. In these situations, our approach is
to keep a copy of the old model so informational content is not lost while moving
forward with the new model.
      A lot of times in this book I’ll tell you what to do with a fair amount of exac-
titude. I apologize in advance for the knee-jerk imperative. It’s not that I assume
my approach is superior to the many other paths to the dollar value of the asset.
Given the organic nature of the process I’ve built, it works well in its totality but is
                                                                 Introduction   •   xix



unreliable in its bits and pieces. And I can’t put please or kindly before every com-
mand; all that wheedling would eventually get on your nerves too.
      This book sets out to accomplish much, and to get it all done, we need to
stay on task. The operative metaphor I’ll sometimes use in this process is the
cattle drive. Yes, I’ll pause to spin some stories along the way. I could argue the
merits of every financial theory we encounter until the campfire goes out. But
the imperative is to push those cattle a little farther down the trail every day.
      Sticking with the western theme, old analysts can come to resemble old
cowboys left too long in the sun: similarly grizzled and at risk for turning crusty,
curmudgeonly, and cynical (and we’re not even out of the C’s). Despite the inevi-
table crankiness that sets in with too much time in the Wall Street sun, I’ll try to
keep my rants to a minimum and mainly stay on topic.
      Within every financial writer lives the soul of a frustrated novelist. Long
before learning the basics of finance, such writers learn the basics of three-act
drama and story arc: the setup, the “backstory,” the denouement, and so on.
Central to delivering an effective and satisfying conclusion at story’s end is the
resolution of those themes identified within the dramatic exposition. These back-
story elements can turn the early going into a tough slog, but they can equally
make the climactic wrap-up all the more compelling and satisfying.
      In this book, the income statement serves as a species of dramatic exposi-
tion, and valuation technique as the story arc. I’ll begin to weave all the strands
in the discussion of “Industry Matrix.” And with “Peer Derived Value,” we reach
a climax as all the elements of our prior work coalesce to enable a new and pro-
prietary valuation technique. I offer final thoughts on “Dollar Value of the Asset”
as a postscript. If you keep that in mind, the detailed slog through the income
statement with which I begin the book may not seem so arduous.


The Uses of Modeling
If you’ve already plunked down money for this book, it’s a bit late to ask this
question, but I’ll ask it anyway: why model? It may seem like needless trouble in
an era in which historical and forecast data is widely available. I’d suggest that if
you haven’t run this data through your fingers, it can be more misleading and
dangerous than no data at all.
      I’ve tried to create a concise and sufficiently compact model that with time
you’ll be able to build and populate in little more than a day. While even that may
seem like too much of a commitment in your busy life, the asset manager often
finds that key names—Kimberly-Clark, or IBM, or Emerson—will be bought
xx   •   Introduction



and sold numerous times in the course of your investing life. Doing the modeling
process yourself rather than relying on outside sources makes you better able to
wring value from what you read in the 10-K or 10-Q. Throughout this book, I’ll
draw on my extensive modeling experience in the communications technology
space while straying now and again into semiconductors, manufacturing, and
other industries. I always try to use illustrations with bearing across the entire
universe of investable equities (banks excluded; their income statement presenta-
tions warrant their own book).
      Modeling will fine-tune your BS detector. Time and again you will hear
CFOs and CEOs promise millions of dollars in savings from this or that restruc-
turing initiative. With the market’s short memory, management is rarely held
accountable for failure to deliver on these promises. Investors are too busy chas-
ing the next carrot of operating cost reduction down the road to notice when they
are whacked with the stick of earnings shortfall. But the careful modeler will have
the quarterly operating cost totals before his or her eyes; and they’re visceral,
because he or she has typed them in.
      The financial models referenced in this book were built over years and in
some cases decades. The mature models tend to be hundreds of columns wide
and hundreds of rows deep. Presenting these models “as is” is simply impractical
on the page. The book includes upward of 65 figures or examples that represent
snapshots or snippets in each case of a living model. In constructing the exam-
ples, I faced a choice: freeze the snippets to include the original column and row
references; or use the column and row references created in the scaled-down
snippets. Had I used the first choice, an income statement example might have
referenced cells within columns BF through BJ, and rows 167 through 174. I felt
this would be needlessly confusing. Accordingly, throughout the text I use col-
umn and row references created within the snippets. Because putting column
and row headers on the snippets would give a false impression of the size and
scale of the actual models, I have eliminated column and row headers from the
figures. As much as possible I have indicated truncation within the models by the
use of dark shading. Light shading is used to highlight rows, columns, or cells of
interest.
      Finally, a word on nomenclature, or specifically the pronouns, that are
employed throughout this book. As you may have noticed, I’ll variously use I or
we. This is not as random as it may at first appear. The word I refers to the author,
Jim Kelleher, and generally relates to my anecdotal experiences in some area. The
word we refers to the legions of analysts, investors, students, and others who have
directly and indirectly contributed to this process; to them I am eternally
grateful.
                                                             PART 1

           INCOME
           STATEMENT
           PRESENTATION



Overview
A successful, comprehensive financial workbook that includes modeled financial
statements as well as valuation enablers is so neatly shaped that, for the model
builder, it is hard to know where to begin. In classic chicken-or-egg fashion, we
ask: should we start with financial modeling or with valuation?
      Any discussion of beginnings begins with “Begin the Beguine,” a 1940s pop
song based on the eponymous dance that is a close cousin to the rumba. Imported
from Martinique and Guadeloupe, the beguine is a slow dance requiring very
close partners to move tightly in sync. Our goal is to make modeling and valua-
tion move in sync.
      In the business school library, you’ll see row upon shining row of tomes
dedicated to the topic of valuation. Squeezed in at the end, maybe, will be a single
volume on financial statement modeling. And why should it get any more shelf
space, given that income statement modeling is treated as a straightforward exer-
cise in percentage-of-revenue analysis?
      In the dance of asset analysis, valuation is Fred Astaire while modeling is
Elaine from Seinfeld. The academic treatment of valuation is elegant and compre-


                                                                                   1
2   •   Income Statement Presentation


hensive, while the treatment of modeling is sometimes slapdash and awkward.
Yet how can the most precise and imaginative valuation techniques be put to full
use if the modeling of financial statements is treated as an afterthought? We’re
going to move forward in this book with the premise that financial statement
modeling is Ginger Rodgers to Fred Astaire’s valuation: doing everything back-
ward and in high heels, and always in tight coordination.
      In our process, we build what we call the income statement presentation in
four stages or phases. Chapters 1 and 2 introduce Phase 1, which we call “The
Income Statement,” and is dedicated to building a flexible and responsive model
of the income statement, along with a “percentage-of” (mainly revenue) section
immediately thereafter. Phase 1 is by far the longest section, so I have broken it
up into two chapters to ease your digestion of lessons learned.
      Within Phase 1, we have several tasks that should be performed consecu-
tively. We will create a visually appealing and informative format. We will gather
historical data for past years and quarters. We will adjust the historical presenta-
tion to include interim periods. We will next adjust historical periods to accom-
modate real-world events, such as the one-time or nonrecurring costs that
influence adjusted or non-GAAP earnings. Next, we will build an eight-quarter
forward income statement model that accommodates both a GAAP and an
adjusted representation of earnings. We will build this model even for companies
that only report GAAP results, based on the reality that cyclical impairment of
assets eventually leads every company to be valued on adjusted results from time
to time.
      Phase 2, called “Segment Modeling,” is presented in Chapter 3 and is based
on presentation of company-issued data that we will use to model the consoli-
dated revenue line. In Phase 3, located within Chapter 4, we discuss modeling
segment operating income, mainly so we can use it in a proprietary technique.
In a variation on standard percentage-of-revenue modeling, we will demonstrate
how to model so-called percentage of difference (between revenue and segment
operating income) up from the segment level into the P&L.
      Chapter 5 details Phase 4, called “The Worksheet,” and represents the mar-
shaling of company reported and anecdotal information, public information
from various sources, purchased information, and other data to model the indi-
vidual revenue segments. At this point we also discuss tailoring the cost inputs,
as well as replicating the model for future years.
      In Chapter 6, we consider several special challenges through the prism of
one company: Ericsson. Consideration of this company’s experience in the U.S.
market enables us to address multiple issues, such as ADR-to-stock equivalency
                                                  Income Statement Presentation   •   3



and changes in that relationship; stock splits; joint venture modeling; and other
special circumstances. And in Chapter 7, we round out our modeling tool-kit
with an introduction to normalized earnings as well as the OLS (ordinary least
squares) refinement to determining compound annual growth rates.
       We begin at the beginning with the income statement because the modeling
of per-share earnings drives so much of the valuation process and figures so
heavily in the market’s valuation mindset. Within a standard Excel workbook, on
worksheet 2 and in cell A1, write “[Company Name] Income Statement.” On the
identity tag on the bottom of the worksheet, we title this page “Incm Stmnt.”
       We always begin with the income statement in our models because an inor-
dinate portion of valuation is derived from one simple metric: the price/earnings
(P/E) multiple. It’s easy to understand the lure of the P/E if you think like an asset
manager, particularly all those generalists out there charged with making money
on the funds entrusted by friends, family, and a group of clients. P/Es are easy to
understand, apply to every equity, and lend themselves to instant analysis: that
is, historical P/Es are readily available for comparison to projected P/Es.
       We’ll develop and analyze many more valuation methodologies across the
course of this work, but we are keenly attuned to the importance of accurate
modeling of earnings. Keep in mind that “earnings” bear no relation to the cash
generated by a company in the course of its everyday business.
       Analysts recognize earnings as a witch’s brew of compromises and uncer-
tain inputs that can be subject to, if not manipulation, then at least massaging.
These inputs include revenue recognition, inventory recognition decisions,
straight-line depreciation, “accounting”-based as opposed to cash-based taxes
and interest. Earnings per share (EPS) are further complicated by changes in the
share base. To arrive at diluted EPS, net income is divided by a share base that
rises and falls on numerous inputs, including net income level and the stock price
in relation to the status of various common stock equivalents.
       Yet earnings are not just universally accepted; they are a bedrock of valua-
tion, and for one key reason: they are calculated the same way. The constancy of
earnings brings us to an important early takeaway. Sometime the value of valu-
ation is not in its accuracy but in its constancy. Because earnings are always real-
ized in the same fashion for every company, they form a common ground for
analysis—even if that common ground is tilted by all the inherent uncertainty
in the inputs.
       In Russia, before embarking on a long journey, it is customary for travelers
to sit a few moments in quiet contemplation. (In the United States, we pause only
4   •   Income Statement Presentation



long enough for, “Where are the car keys?”) Let’s take a moment to contemplate
the task ahead. Broadly, we’re going to wade into the digital data stream to seek,
capture, and tame data for use in deriving equity value. Specifically, we’re going
to model financial statements, use modeled and historical data to value equities,
and replicate the process to further enhance individual equity valuation within
a group of like companies. Let’s get started on a vital task for any analyst: accu-
rately modeling the income statement.
                                                   Chapter 1
           PHASE 1:
           INCOME STATEMENT
           AND MARGIN MODEL,
           PART 1




Modeling in a Workbook
Modeling requires populating countless cells on a spreadsheet. Some of those
cells hold complex formula; others hold numbers. As you build your model, you’ll
know which is which; but to the uninformed eye, a formula cell and a number
cell are indistinguishable.
       One difficulty with describing a dynamic process in a static medium such
as a book is that our spreadsheet examples (designated as Figure 1.1, Figure 1.2,
etc.) cannot fully display the formulas residing within; that would make for inde-
cipherable gibberish. We thus ask that you read the text closely, refer carefully to
the model when requested, and make the needed abstract connections between
written word and illustration on the page.
       We are going to conduct our analysis of a single equity, as much as possible,
on a single workbook with linked worksheets. Later we’ll discuss a matrix work-
book in which we aggregate data from the individual workbooks. Later still we’ll
throw data from the matrix workbook back to the individual worksheet to pro-
vide a kind of peer group relative ranking that will help inform the valuation
calculation. For now, though, we can begin with a focus on the individual equity
workbook.

                                                                                   5
6   •   Income Statement Presentation



       Our goal is to build a workbook that, once created, can be replicated for
other companies in the coverage universe with the least amount of disruption.
We’ll spend a great deal of time discussing communications equipment compa-
nies, so let’s borrow a metaphor from the industry to describe a key element of
our task. In an astounding transformation, the legacy communications infra-
structure has been torn down and replaced with protocols such as Ethernet bor-
rowed from the data network. Ethernet was chosen partly because it is ubiquitous
(there is only one Ethernet jack in use the world over) but mainly because it is
easily replicable.
       When we build a workbook, we’ll have a few tasks in mind. We want the
individual equity workbook to fully capture value for a single stock. But we also
want it to lend itself to relatively straightforward manipulation whereby it can
transformed into a second company model, and a third, and so forth.
       No two companies are alike; equally important, no two companies report
their financial data in exactly the same way. Creating a replicable workbook is a key
element of our task, given that we are working in a world of jangling phones and
deadlines, and we certainly are not within that proverbial ivory tower—which, as
metaphors go, must be at least as politically insensitive as skinning cats.
       Our individual company workbook will have several worksheets. At a mini-
mum, our generic workbook will include income statement (or income statement
presentation, as we’ll call it); ratios and valuations, where we’ll conduct historical
comparable valuation and also aggregate annual ratios and annual financial
statements; a present value worksheet for discounted free cash flow valuation; a
query page for real-time pricing; and a worksheet to calculate “smoothed” his-
torical growth rate and forecast normalized earnings. In practice, any workbook
used to model and value an individual equity will over time sprout additional
worksheets that will be used to track changes in manufacturing footprint, model
the combination of acquired assets, or accommodate other real-world events.
       We’ll discuss formatting of individual worksheets and necessary historical
backup data as we go. For the past two years, we have been modeling in .xlsx
workbooks. All our techniques work in prior-generation .xls workbooks as well.
If you are new to this process, we highly recommend that you begin with the .xlsx
workbooks consistent with the latest version of Excel.
       Given that we need to create a replicable workbook, choose for your first
model the most “normal” company in coverage. In other words, for your tem-
plate use a company with steady and even boring growth, consistent profitability,
little to no history of changing segment presentations, and a reasonably stable
business model. If no such company exists in your coverage, take the best of the
bunch.
                            Phase 1: Income Statement and Margin Model, Part 1   •   7



      Every workbook needs to access a source for real-time pricing of the asset.
As a first step, we will label worksheet 1 as our “Query” page. Our recommended
price-data source, and the one used in all our workbooks, is the Money page from
MSN. Guidelines for linking data to this worksheet will vary; the process typi-
cally requires that a ticker be placed on the query worksheet for immediate refer-
ence, or on an adjacent worksheet for a linked reference. In addition to the ticker
for the individual equity, we will need real-time pricing for the market bench-
mark—in this case, the Standard & Poor’s 500 Index (S&P 500). Given that the
pricing function can accommodate dozens of tickers, it is also highly useful to
include prices for the other major indexes [Dow Jones Industrial Average (DJIA),
Nasdaq, and New York Stock Exchange Composite (NYSE Composite)]; you may
also want to include industry indexes in which the equity is a member.
      Figure 1.1 shows the Query worksheet in the Motorola (MOT) model. In
addition to prices for MOT, we have priced the S&P 500; much further down the
line, this will come in handy for calculating the relative P/E. Also note what is
not there; at this point, we have no need to price the peer group.


Tracking Down Historical Data Sources
Having built several hundred income statement models, I can tell you that indi-
vidual company income statement presentations are like snowflakes or finger-
prints: remarkably similar, but if you squint, you can see the differences.
Individual income statements aren’t issued by cookie cutters; they come from the
finance departments of disparate companies, peopled with individuals with their
own quirks.
      The modeled income statement has at its heart a variety of percentage-of-
revenue inputs. The trouble is, being married to pure percentage-of-revenue pro-
cedure robs your model of nuance and precision. As we’ll demonstrate later on,
depending on the data, you can build an income statement model not on percent-
age of sales but on percentage of difference between forecast revenue and forecast
operating income.
      To model future financial statements, you need historical data as issued by
the company. This historical data will be most usefully recorded in the company’s
own filings with the Securities and Exchange Commission (SEC), available on
every company Web site (typically, see Home Page>Investor Relations>Fi-
nancials>SEC filings). A variety of pay-for sites exist that enable download of
data into Excel; they may charge a one-time fee or a subscription. Because these
come and go, we’ll limit our mentions to EDGAR.
Figure 1.1

A Query page from MSN Money in the Motorola model. Real-time pricing of the asset and its benchmark (the S&P 500) is vital to the valuation process.

Stock Quotes Provided by MSN Money
Click here to visit MSN Money
                                       Previous                                                  %    52 Wk 52 Wk                          P/E      # Shares
                                Last    Close              High        Low Volume Change Change High          Low Market Cap EPS          Ratio       Out
Motorola Inc         Chart News   7.04      7.28              7.29       7.01 5,327,398  -0.24 -3.30%    10.5   2.98 16,159,362,472 -1.87       0 2,295,364,000
S&P 500 INDEX Chart News        997.17   1012.73            1012.6      994.6         0 -15.56 -1.54% 1303.04 666.79              0     0       0             0

Symbol Lookup              MSN Money Home                            Microsoft Office Tools on the Web
Find stocks,               Discover MSN Money’s                      Get the latest from Microsoft Office
mutual funds,              tools, columns, and more!
options, indices,
and currencies.

Terms of Use. © 2009 Microsoft Corporation and/or its suppliers. All rights reserved.
Data providers
Canadian investment fund pricing (c) 2009 CANNEX Financial Exchanges Limited.
Copyright © 2009 Thomson Reuters. Click for Restrictions.
Analyst Recommendations data provided by Zacks Investment Research.
Fund data provided by Morningstar, Inc. © 2009. All rights reserved.
Quotes supplied by Interactive Data.
StockScouter data provided by Gradient Analytics, Inc.
                           Phase 1: Income Statement and Margin Model, Part 1   •   9



       Prepare for some confusion. EDGAR (electronic data gathering, analysis,
and retrieval) is a useful and free service sponsored by the SEC that simplifies
the tracking down of SEC filings for every company. News headlines on the
Web, however, will frequently steer you to EDGAR Online Pro, which is a pay
service. Your particular situation—including the extent of your coverage
responsibilities, your company’s resources, and time constraints—will deter-
mine whether you rely on a vendor for historical data or transcribe the data the
old-fashioned way.
       Public companies in the United States are compelled to report their results
every 90 days. They may or may not do so through their investor relations func-
tion. Some firms provide all the bells and whistles, including 20-page press
releases, an accompanying slide deck, conference call, and even conference call
transcript. Sometimes you get no more than a press release with abbreviated
financial data. More and more companies have discovered the virtues of being
user-friendly toward investors. As part of that outreach, many firms now furnish
their financial statements in Excel format for easy download into your model.
       You will also find that financial data is available on the major financial
sites, both free (Yahoo! Finance, Google Finance) and pay (Bloomberg). True,
you can gather financial statements from these and other sources. But be careful;
you don’t want a cookie-cutter synopsis of all the main points. That is typically
what you get on Yahoo! Finance, which uses a cooked-down version of the income
statement, balance sheet, and cash flow statement.
       For your sanity while maintaining a model over the long term, you’ll need
to model every line, because you need your model to think and act like the issu-
ing CFO. So here is another of our bona fides: we will begin by recreating the
model exactly as the issuing company does. Later we will change it to accom-
modate the real world as well as our own organizational needs.
       We recommend beginning with (at least) a base of the five preceding his-
torical years, with the immediately preceding two years modeled in historical
quarters. Perhaps most important, for all annual periods, we are going to reinvent
the past one or two years not as a series of discrete quarters, but as a process.


Amending the Status Quo
To my thinking, the standard business school guidance on income statement
modeling is replete with problems. The standard guidance is something like this:
Represent the figures in the income statement as percentages of revenues. Line
the quarters up one next to the other; sum them for the full year. For certain
10   •   Income Statement Presentation



high-priority inputs, position year-over-year or quarter-over-quarter percentage
changes immediately below the line item. Situate the percentage of revenue per-
centages at the bottom of the sheet for each quarter; link them to the individual
cells below the revenue line. Rinse and repeat for year 2.
      Some form of percentage-of-revenue modeling is a given, and we’ll use per-
centage of revenue as a starting point. But, to my thinking, the importance of
presentation on the page is overlooked in that standard template. So too is con-
venience, a vital need for the time-challenged analyst. For example, Excel makes
for easy summing. The introduction of annual or sequential rate-of-change per-
centages in the column thwarts that process or introduces inaccuracy into the
model. Sure, percentages amount to less than 1, but why introduce any inaccu-
racy in a carefully modeled income statement?
      Here’s another early key takeaway: think like the reporting company. The
reporting company sees the year not as four discrete baskets but as a process.
Companies are dynamic; they set goals based around new product ramps and
extensions, cost management, competitive progress, and host of other things.
Therefore, our income statement mirrors the company’s presentation and models
the year cumulatively. Moreover, the progress or lack thereof should be visually
immediate to the modeler. To this end, we include summary columns as the year
progresses, for the first half, nine months, and full year.


Structuring the Income Statement Model
If you are beginning with that standard model of four side-by-side historical quar-
ters from the preceding year lined up next to one another, adjust them as follows:
Immediately to the right of historical quarter 1 and between quarter 2, add a col-
umn enabling visualization of year-over-year percentage change for every line item.
Italicize this column too, further distinguishing it from any column containing
dollar values. Next comes historical quarter 2; similarly position next to it a column
enabling year-over-year percentage change in every line item.
       Here’s where our divergence begins, enabling us to see the year as a process.
In the next column we sum the half-year to date. Then comes quarter 3 and its
attendant annual percentage change column, followed by a nine-month sum of
the year to date. Finally we position quarter 4 and its attendant percentage change
column, with the full-year summary and its annual change column.
       In Figure 1.2, we see the four historical quarters for ADTRAN in 2004. I
do not typically include a year-over-year rate of percentage change comparison
column for the half year or the nine months, because investors don’t usually
Figure 1.2

The four-quarter and interim-period P&L presentation for ADTRAN for the year 2004. The pro forma (PF) line items reflect the company’s conventions for adjusting
GAAP earnings to exclude nonrecurring events and noncash items.

ADTRAN
Income Statement               1Q04    Yr/Yr %       2Q04    Yr/Yr %       1H04       3Q04    Yr/Yr % 9mos04           4Q04     Yr/Yr %      2004    Yr/Yr %
 Net Sales                       114.0     32%         120.6     33%         234.6      115.3       9%   349.9           104.9       -8%       454.8     15%
 COGS                             48.5     24%          50.8     25%          99.4       48.5       3%    147.9           45.7       -4%       193.5     11%
 Gross Profit                      65.5     39%          69.8     40%         135.3       66.8     13%    202.0            59.3      -11%       261.3      18%
 SG&A                             22.3     11%          24.0      18%         46.3       22.2       8%    68.4            23.4        6%        91.8      10%
 R&D                              14.8       4%         15.9     17%          30.7       18.9     26%      49.6            17.8      17%        67.4      16%
 Stock Option Expnsng
 Operating Costs & Exp              37.1       8%        39.9       17%        77.0       41.1       15%       118.0       41.2      10%       159.2          13%
 PF Optng Costs & Exp
 Operating Earnings                 28.4     125%         29.9      90%        58.3      25.7        10%        84.0       18.1      -38%       102.1         27%
 PF Oprtng Ernngs                   28.4     125%         29.9      90%        58.3      25.7        10%        84.0       18.1      -38%       102.1         27%
 Interest Income                   (2.2)                 (2.6)                (4.8)      (2.1)                 (6.9)      (3.8)                (10.7)
 Interest Expense                    0.6       -4%         0.6      -2%         1.3        0.6        3%         1.9        0.6       2%          2.5          0%
 Other Income                      (0.4)      -93%           -                (0.4)          -                 (0.4)          -                 (0.4)        -78%
 Net Realized Invstmnt Gains         0.1   -9609%            -                  0.1          -                   0.1          -                   0.1

 Pretax Income                     30.2      104%        31.8       77%        62.0       27.2        6%        89.2       21.2      -30%       110.4         24%
 PF Pretax Income
 Income Taxes                       9.8      128%        10.3       85%        20.2        8.5       -1%        28.7        6.4      -28%       35.0          28%
 PF Income Taxes
 Tax Rate                          32%        12%        33%         5%        33%        31%        -6%       32%       30.0%        2%        32%            3%
 PF Tax Rate
 Net Income                        20.4       94%        21.5       73%        41.9       18.7        9%        60.6       14.9      -31%       75.4          23%
 PF Net Income
 Basic Shares Outstndng            79.5        4%        79.3        4%        79.4       77.8        0%        78.9       76.6       -3%       78.2           2%
 Diluted Shares Outstndng          82.8        7%        82.2        3%        82.5       80.4       -1%        81.8       78.7       -5%       81.0           0%
                                                                                                                                                        (continued)
Figure 1.2 (continued)


ADTRAN
Income Statement         1Q04    Yr/Yr %   2Q04    Yr/Yr %   1H04      3Q04    Yr/Yr % 9mos04     4Q04    Yr/Yr %    2004    Yr/Yr %
Reprtd EPS Basic            0.26     87%      0.27     66%      0.53      0.24       8%    0.77      0.19     -28%      0.96     21%
 Reprtd EPS Diluted         0.25     82%      0.26     68%      0.51      0.23      10%    0.74      0.19     -27%      0.93     22%
 Adjstd EPS Basic           0.26     87%      0.27     66%      0.53      0.24       8%    0.77      0.19     -28%      0.96     20%
 Adjstd EPS Diluted         0.25     82%      0.26     68%      0.51      0.23      10%    0.74      0.19     -27%      0.93     22%
 Dividend                   0.08              0.08              0.16      0.08             0.24      0.08               0.32
                           Phase 1: Income Statement and Margin Model, Part 1   •   13



think in these terms. But these interim-period columns are available for easy
comparisons.
      Modelers may wonder: where are the sequential comparisons? These can
matter more than annual comparisons for some companies, particularly firms
with a limited seasonal influence and those still in their early growth phase.
Given proximity on the page, visual rather than percentage change is sufficient
for sequential comparison for many line items. For higher-priority inputs, we will
furnish formal sequential comparisons as needed, mainly further down the
income statement presentation.


A Not-So-Random Walk Down the Income Statement
For thoroughness, let’s take a quick walk down the income statement. Purists, be
forewarned: the emphasis here is on quick, and what follows is by necessity a very
cursory treatment of complex topics. Much more thorough treatments of income
statement line items are available elsewhere. Whole volumes have been devoted
to revenue recognition, depreciation, FIFO-LIFO accounting, tax treatment, and
other issues; it would be a distraction to detail them here. At least on a summary
basis, the model builder must know the basis of individual line item inputs in
order to best represent them.
       The first line item on the income statement is revenues. Many companies
have taken to breaking out product from service revenues. We always welcome
enhanced granularity.
       Remember that revenues do not reflect cash that has ca-chinged the cash reg-
ister within a 90-day span. Instead, it reflects revenues “recognized” in a period.
Given the payment terms of 30 to 60 days in the United States and often much
longer overseas, recognized revenues may not jingle the cash register for months.
This tally also includes proportional payments from lump sums collected earlier
(e.g., one-twelfth of a 12-issues-per-year magazine subscription). Service revenues
in particular will reflect timed recognition of cash paid either in lump sum or over
a period for regularly scheduled maintenance and/or special services.
       Cost of goods sold, or COGS, primarily reflects what is recognized (again,
as opposed to what is spent) to physically produce goods or, in some cases, ser-
vices. This mainly includes costs for materials, manufacturing, maintaining the
supply chain, procurement, and most of the costs of keeping the factory going.
The difference between revenue and COGS is gross profit or gross margin.
       Another element in COGS is depreciation, reflecting assumed loss of value
in a physical asset from usage and time’s passage. Formerly, depreciation was
14   •   Income Statement Presentation



lumped in DDA (depreciation, depletion, and amortization); depletion is mainly
of interest to energy investors, while companies now prefer to separately list
amortization (more below in our continuing operations discussion).
       Throughout this text, when we address financial statement accounting, our
shorthand is “book” accounting; what happens in the real world we refer to as
“cash” accounting. On a book basis, assets are depreciated on a straight-line basis,
meaning in equal increments over the asset’s useful life. In the real world, com-
panies use the more tax-friendly accelerated depreciation, where high levels of
depreciation are recognized early on followed by ever-decreasing decrements.
This discrepancy is one of several that influence the difference between income
statement accounting and real-world accounts.
       COGS will also be influenced by inventory treatment decisions. LIFO
accounting is an acronym for “last in, first out,” meaning a company records a
profit from sales based on the company’s cost for its most recently purchased inven-
tory. Remember that this is not a product-flow issue but an accounting recognition
one. LIFO is regarded as better matching costs to prices. In a period of normally
increasing prices, or inflation, LIFO reduces profits and therefore taxes compared
with FIFO, which is an acronym for “first in, first out.” FIFO is considered appro-
priate when a company ships an undifferentiated product, such as a commodity.
Start-ups may favor FIFO; in a period of normally increasing prices, FIFO increases
inventory values and makes a company look “bigger.” For the small company, con-
veying size can matter in winning new financing. The prevailing International
Financial Recordings Standards, or IFRS, does not accept LIFO.
       Continuing down the income statement we come to the two or three key
elements of operating cost. Most often, these are (1) research and development
(R&D) and (2) selling, general, and administrative (SG&A); less often, they are
(1) R&D, (2) selling and marketing (S&M), and (3) general and administrative
(G&A). Fans of the granular, we always prefer the second three-part representa-
tion, because it better breaks out selling and marketing costs from the cost of the
corporate function.
       These key operating costs are also of interest because embedded within
them are stock-compensation costs that some companies, in league with their
covering analysts, will disaggregate in order to better represent real costs of doing
business. FAS 123R is a relatively new standard that compels corporations to
reflect components of stock-option compensation within the income statement
line item in which they are expensed. In other words, if a portion of an engineer’s
compensation is in the form of stock, this payout must be included as a cost
within R&D. Prior to this standard, the cost to the company of stock granted to
that engineer would not appear in the R&D line item. FAS 123R has had a pro-
                           Phase 1: Income Statement and Margin Model, Part 1   •   15



found impact on operating costs and thus on net income. Its disruptive appear-
ance in current income statements makes a strong argument for staying current
on accounting standards, and on creating a financial model flexible enough to
accommodate them as they occur.
       Other elements of operating costs that frequently earn a regular spot as
separate line item on the income statement include amortization of intangibles
and “other,” a deliberately vague catch-bag of items that may include legal settle-
ments, disposition of minor assets, current exchange effects, or “other.”
       In an ideal world, companies would operate flawlessly, every acquisition
would be perfectly timed and seamlessly integrated, and the economy would
always be in the up part of the cycle. Back here in the real world, companies
overshoot on expectations and underdeliver on performance, buy ill-fitting assets
that gum up their core operations, and navigate a macroeconomy in which bumps
alternate with pitfalls.
       The space on the income statement between the main operating cost items
(i.e., R&D and SG&A) and GAAP operating income is clogged with monuments
to ill-planning and economic reality, mixed in with noncash items that must be
recognized in the course of normal business. Companies all recognize these dif-
ferently. Some larger companies may fold their mistakes into a larger operating
cost line item such as SG&A, while smaller companies that need to put the best
spin on their performance will exclude every one-time or noncash item to bolster
earnings. These items mainly include restructuring, asset impairment, goodwill
impairment, acquisition integration, write-down of acquired in-process R&D,
settlement costs, and other. (Note that while actual asset-acquisition costs do not
flow through income statements, costs attendant on that acquisition or its inte-
gration can be included.)
       The difference between gross profit and the sum of operating costs is oper-
ating income. Operating income is then impacted primarily by interest costs and
interest income. Depending on presentation style, this may be reflected as sepa-
rate line items or a net figure. Again depending on style, companies may choose
to include line items that typically recognizes value changes in investment assets,
in shares held in trust for a settlement, or other similar items that impact account-
ing net income but are deemed to be nonoperating in nature. Interest earned on
cash and equivalents and investments, as expressed on the income statement,
often has a close correlation with real-world cash from these items. By contrast,
“book” or income statement interest cost can vary, sometimes meaningfully,
from cash interest costs.
       Operating income less these items is pretax income. Income taxes on a book
basis may look like a single number, but they are actually an amalgam of inputs
16   •   Income Statement Presentation



from different jurisdictions. Corporate income tax rates for industrialized nations
tend to run in the 30% to 40% range; emerging economies are usually character-
ized by low income tax rates meant to encourage new business investment. Given
the widely varied tax rates from different nations, a globally diverse company
may report an operating loss but still be liable for net income taxes because the
company made money in one or more countries despite losing money in others.
The difference between book taxes and cash taxes from varying jurisdictions and
(reaching back to our discussion of depreciation) the difference between straight-
line and accelerated depreciation are among several items that can cause a pro-
found difference between book taxes and cash taxes.
      From pretax income on down, individual company presentation of the
income statement can take on a very individualized tint. The main contribu-
tor to this variation is the relationship with partner companies. On a very
cooked-down basis (purists: shield your eyes), and allowing exceptions, joint-
venture (JV) companies are consolidated on the income statement of the
majority partner (i.e., the one that owns more than 50%). Consolidation
means that all the joint venture’s revenues are recorded in one (but not both)
of the partners’ top lines. For 50/50 joint ventures, one partner elects to con-
solidate and the other doesn’t; there is no joint consolidation. While many
companies, including nearly all large-cap companies, participate in JV com-
panies, many do not.
      Nonmajority partners, particularly those owning 20% to 50% stakes in the
joint venture, will typically record proportional income (or loss) as equity income
below the pretax income line and above net income. Majority partners will rec-
ognize as a proportional cost (or gain) against JV profits (or losses). The majority
partner is effectively disgorging profit (or loss) to its smaller partners propor-
tional to the position the majority partner does not hold in the joint venture.
Large corporations may be majority partners in some joint ventures and minority
partners in others; their EI (equity income) and MI (minority interest, also called
noncontrolling interest) lines are an amalgam of these inputs. In Figure 1.3, we
see Corning’s equity income and minority interest line items for two years (2002
and 2008) with two very different profiles.
      A less-frequent variable standing between pretax income and net income is
under the heading “Accounting Changes.” These can be a one-time impact on the
income statement from a change in accounting treatment mandated by the
Financial Accounting Standards Board (FASB).
      A final adjustment can and frequently does intrude. If a company indicates
plans to sell an operating asset, net results from that asset are stripped from con-
solidated operations and moved to discontinued operations. This accounting
Figure 1.3

Within the (compressed) income statements for 2002 and 2008, we show a comparison of Corning’s equity income and minority interest. Note that minority inter-
est, nearly equal to equity income in 2002, had all but vanished by 2008. Meanwhile, equity income had grown so large as to become roughly equivalent to operat-
ing income.

Corning Inc.
Income Statement              1Q02 2Q02 1H02 3Q02 9mos02 4Q02                       2002          1Q08 2Q08 1H08 3Q08 9mos08 4Q08                           2008
 Revenues                         730      753   1,483     762    2,245        736     2,981       1,617.0 1,780.0 3,397.0 1,555.0 4,952.0 1,084.0           6,036.0
 Cost of Sales                    655      643   1,298     633     1,931       631     2,562        773.0 858.9 1,631.9      820.0 2,451.9   777.0           3,228.9
 Gross Profit                        75     110     185     129       314       105        419       844.0    921.2 1,765.2   735.0 2,500.2   307.0           2,807.2
 SG&A                              188     188     376     157       533       183        716       242.0    249.2   491.2   220.0    711.2  179.0             890.2
 R&D - Engnrng                     126     131     257     113       370       113       483         151.0   173.6   324.6   160.0   484.6   153.0             637.6
 Amortization Intngbls              11      11      22      11        33         10        43          2.0     5.0     7.0     2.0      9.0    5.0              14.0
 Acquisition, Asbsts                 -       -       -       -         -          -         -       (327)        - (327.0)       6 (321.0)    (28)           (349.0)
 Restructng                          -    (68)    (68)    (22)      (90)      (86)     (176)                     -             (2)    (3.0)     22              19.0
 Impairment                                494     494     125       619     1,461     2,080
     Operating Earnings        (250.0) (646.0) (896.0) (255.0) (1,151.0) (1,576.0) (2,727.0)          777.0    493.4 1,270.4     349.0    1,619.4    (24.0) 1,595.4
     PF Oprtng Ernngs          (250.0) (220.0) (470.0) (152.0) (622.0) (201.0) (823.0)                451.0    498.4 949.4       355.0    1,304.4    (25.0) 1,279.4
 Interest Income                  (14)    (10)    (24)    (10)      (34)        (7)      (41)        (30.0)   (36.7) (66.7)     (22.0)     (88.7)    (11.0) (99.7)
 Interest Expense                   48      44      92      44       136         43       179          18.0     20.1    38.1      15.0       53.1      11.0    64.1
 Debt Prpmnt
 Other, Net                           9       -       9       1         10       28        38                 (40.0) (41.0)        30.0     (11.0)     32.0      21.0
     Pretax Income              (293.0) (680.0) (973.0) (290.0) (1,263.0) (1,640.0) (2,903.0)         790.0    550.0 1,340.0      326.0   1,666.0    (56.0) 1,610.0
     PF Pretax Incm               (293)   (186)   (479)   (165)     (644)     (179)     (823)         464.0    555.0 1,019.0      332.0   1,351.0     (57.0) 1,294.0
 Incm Tax                          (50)   (184)   (234)    (91)     (325)     (401)     (726)          66.0     82.5   148.5     (60.0)       88.5   (23.0)     65.5
 PF Incm Tax                       (50)    (47)    (97)    (45)      (141)     (48)     (189)          67.3     80.5    147.8    (15.0)     132.8    (18.0)    114.8
 Tax Rate                         17.1% 27.1% 24.0% 31.4%          25.7%     24.5%     25.0%          8.4%    15.0% 11.1%       -18.4%       5.3%    41.1%     4.1%
 PF Tax Rate                      17.1% 25.0% 20.1% 27.0%          21.9%     27.0%     23.0%         14.5%    14.5% 14.5%        -4.5%       9.8%    31.6%     8.9%
 Income bfr MI & EE               (243) (496)     (739)   (199)     (938) (1,239) (2,177)              724      468     1,192      386      1,578      (33)    1,545
 PF Income bfr MI & EE            (243)   (140)   (383)   (120)     (503)     (131)     (634)           397      475      871       347      1,218     (39)     1,179
                                                                                                                                                         (continued)
Figure 1.3 (continued)


Corning Inc.
Income Statement                 1Q02       2Q02       1H02       3Q02       9mos02 4Q02       2002       1Q08       2Q08   1H08 3Q08 9mos08 4Q08 2008
 Minority interst                       6          6        12           5          17      81     98.0          1          -    1.0   -    1.0   -      1.0
 Dividends Cnvrtbl Prfrd (Sub)                               -                       -                -                            -          -            -
 Equity Earnings, net Imprmnts       30         25          55        42            97      19    116.0      304        264 568.0    382 950.0  282 1,232.0
 Impairmnt Eqty Invstmnt                                     -                       -                -                            -          -            -
 Incm Bfr Dscntnd Ops              (207)     (465)       (672)      (152)       (824) (1,139)               1,029       732   1,761  768 2,529  249   2,778
 PF Incm Bfr Dscnrd Ops, Prfrd     (207)     (109)       (316)       (73)       (389)     (31)                702       739   1,440  729 2,169  243   2,412
 Dscntnd Ops Income                    8        21          29         19           48    430
 Prfrd Dvdnd                                                        (128)       (128)             (128)
     Net Incm                      (199)     (444)       (643)      (261)      (904)    (709) (1,963)       1,029       732     1,761     768   2,529     249    2,778
     PF Net Income                 (207)      (109)       (316)     (201)        (517)    (31)    (128)       702       739     1,440     729   2,169     243    2,412
 Basic Shrs Out                      945        945        945       1036          977   1230      1039     1569       1572      1571    1574    1572    1576     1573
 Diluted Shrs Out                    945        945        945       1036          977   1230      1039     1604       1607     1606     1575    1595    1577     1591
 Rprtd Basic EPS                  (0.22)     (0.47)     (0.68)     (0.25)      (0.93)  (0.58)    (1.89)      0.66       2.01      1.12   0.49     1.61    0.16     1.77
 Rprtd Diluted EPS                (0.22)     (0.47)     (0.68)     (0.25)      (0.93)  (0.58)    (1.89)      0.64       2.01      1.10   0.49     1.58    0.16     1.75
 Pro Forma Basic EPS              (0.22)     (0.15)     (0.37)     (0.19)      (0.57)  (0.02)    (0.59)      0.45      0.50      0.95    0.46     1.41    0.15     1.56
 Pro Forma Diluted EPS            (0.22)     (0.15)     (0.37)     (0.19)      (0.57)  (0.02)    (0.59)      0.44      0.49      0.93    0.46     1.39    0.15     1.54
 Dividends Declared                                           -                      -                -      0.05      0.05       0.10   0.05     0.15   0.05     0.20
                           Phase 1: Income Statement and Margin Model, Part 1   •   19



treatment need not await final disposition of the asset, or even announcement of
a prospective buyer; it occurs immediately upon determination that the asset is
being shopped, shut down, or otherwise dispatched. The discontinued asset con-
tinues to warrant this kind of treatment as long as it takes for full asset disposi-
tion and related effects, which may straddle several periods.
       Across its reporting history and depending on its business development
strategy, accounting changes, partner relationships, and other factors, every com-
pany will need to represent some basket of these below-the-line line items, includ-
ing equity income, minority interest, accounting changes, and discontinued
operations. The effects of these items can be persistent or transitory, meaningful
or insignificant. Accordingly, no one standard for presentation prevails when a
company is initially presenting its quarterly results.
       A company’s income statement presentation may subtract taxes and as
many as applicable of these items from pretax income and proceed directly to net
income. Sometimes a company may subtract taxes from pretax income to arrive
at net income before MI and EE; present a number; then separately subtract dis-
continued operations and accounting changes to arrive at net income. The point
is, there is no single method.
       Net income is then divided by the share count to arrive at net income per
share, per-share earnings, or EPS; whatever you call it, this is the most intensely
analyzed number coming off the income statement. Basic EPS reflects net income
as is—that is to say, divided by a stated number of shares outstanding. Diluted
EPS features an amendment to both numerator and denominator. The numerator
(i.e., net income) may also include the equivalent of interest that would be earned
by convertible debt, along with other inputs. The denominator (i.e., share count)
will be inflated by various common stock equivalents, or CSEs. These include
various stock options as well as the shares that would be included upon conver-
sion to common of any convertible debt. Investors are most interested in and
make valuation decisions based on diluted EPS.


Nonrecurring and Noncash
Now that you have exactly replicated a company’s single-quarter historical rep-
resentation of its income statement, let’s amend it to make it more useful. Our
first task is to reconcile the real-world reporting with the continuing-operations
analysis that will inform stock valuation going forward. To do so we are going to
add continuing operations lines, which for simplicity and consistency we will call
pro forma (abbreviated PF).
20   •   Income Statement Presentation



      In Figure 1.4, we have shown Cisco’s income statement for 2005 amended
to our presentation style. Like many companies providing both GAAP and pro
forma results, Cisco aids the analytic community by providing a detailed walk-
through between GAAP and adjusted results. That is not the case with some
companies that are coy about adjusted results even though analysts value their
stocks on adjusted results. Refer to this model as we walk through.
      You may or may not be aware that technology companies, biotechs, and
start-ups of every flavor frequently provide two sets of operating metrics for
investor analysis: GAAP results and pro forma results. Increasingly, companies
outside these designated “growth” sectors are providing GAAP and adjusted or
pro forma results as well.
      GAAP results are presented in accordance with an extensive body of
accounting standards prevailing in the home country. Investors are also inter-
ested in how a company is performing on a continuing operations basis, which
is to say excluding one-time items—such as restructuring costs or acquisition-
related costs—as well as noncash items. The list of noncash items can be lengthy;
primary examples include stock option compensation (as outlined in FAS 123R),
intangibles amortization, and write-downs of acquired in-process R&D.
      Although earnings and changes in earnings estimates drive so much of the
investment decision process, you may be surprised to know that no clear consen-
sus exists on whether a company should be modeled on GAAP or pro forma
performance. The decision to model on a GAAP or pro forma basis resides with
the companies that maintain the analysts’ consensus; these companies are to
some degree in league with the analysts. Consensus earnings estimates that
appear on publicly available sites may or may not contain an indication that they
are GAAP or non-GAAP.
      Line items that analysts and investors exclude from pro forma results fall
into two categories: noncash and nonrecurring. Noncash items are those that,
like amortization of intangibles, reflect no real cash cost to the company. But
these items do have accounting effects; for example, amortization causes the
intangibles balance to decline on the balance sheet, and it thus must go through
the income statement. FAS 123R, the provision mandating the expensing of stock
options used as compensation, impacts the cash flow statement.
      Many analysts and investors cast a baleful eye on exclusion of noncash items
from results. The convention was born of the adage “give the little person a
chance.” Exclusion of noncash but nonetheless GAAP expenses was originally
intended for garage-born tech companies and other start-ups. The idea was to
give them time to grow and eventually establish earnings for valuation purposes.
Many of those start-ups have long since burst the garage walls; but, by habit or
Figure 1.4

Cisco provides detailed GAAP and adjusted or pro forma results all the way down its P&L, but it does not publish a P&L in this format in its SEC filings. Modelers
need to incorporate line items for PF (pro forma) values if they wish to capture both GAAP and adjusted earnings calculations in their model.

Cisco Systems
Income Statement                1Q05       Y/Y %      2Q05       Y/Y %      1H05       3Q05      Y/Y % 9mos05           4Q05       Y/Y %      2005        Y/Y %
 Products                         5,033      18.1%      5,106       12.2%     10,139     5,189       9.7% 15,328          5,525      10.3%     20,853       12.4%
 Services                           938      11.9%        956       12.7%      1,894       998      12.1%  2,892          1,056      14.9%      3,948       13.0%
   Total Sales                    5,971       17.1%     6,062       12.3%    12,033      6,187      10.1% 18,220          6,581      11.1%     24,801       12.5%
 Products COGS                    1,646      25.1%      1,669       17.1%      3,315     1,697      16.9%  5,012          1,746      11.0%      6,758       17.2%
 Services COGS                       310     11.1%        340      23.6%         650       355      17.9%  1,005            367      23.2%      1,372       19.0%
   Total COGS                     1,956      22.6%      2,009       18.2%      3,965     2,052      17.1%  6,017           2,113     12.9%      8,130       17.5%
  Total Gross Profit               4,015      14.5%      4,053        9.6%     8,068      4,135       6.9% 12,203          4,468      10.2%     16,671       10.2%
R&D                                  787       7.1%        785       3.4%      1,572       790      -1.4%  2,362            858       9.3%      3,220        4.5%
S&M                                1,102       2.9%      1,132       3.6%      2,234     1,180       4.3%  3,414          1,257       9.3%      4,671        5.1%
G&A                                 225      15.4%        222       13.8%        447       237      10.2%    684            250      25.6%        934       16.2%
Pyrl T+A41x Stk Optn Excrs             1    -50.0%           3     -57.1%          4         3       0.0%      7               5     25.0%         12      -25.0%
Amrtzn Dfrd Stk-Bsd Cmpnsn            40                    39                    79        47               126              39                  165
Amrtzn Intngbls                       60     -3.2%          57     -5.0%         117        54     -10.0%    171              56      -6.7%       227       -6.2%
In-Process R&D                        12      0.0%           2    100.0%          14         6    200.0%      20               6       0.0%        26      766.7%
Acquisition Costs                      -                     -                     -         -                 -               -                    -
    Oprtng Expenses               2,227       5.2%      2,240       4.1%       4,467     2,317       0.2%  6,784          2,471       9.6%      9,255
    PF Oprtng Expenses             2,114      5.6%      2,139       1.2%       4,253     2,207       2.8%  6,460          2,365      10.8%      8,825
    Oprtng Incm                   1,788                 1,813                  3,601     1,818             5,419          1,997                 7,416
    PF Oprtng Incm                1,901      26.3%       1,914     15.9%       3,815     1,928      12.1%  5,743          2,103       9.5%      7,846       15.4%
Gain Sale Invstmnt                     -                     -                     -         -                 -               -                    -
Interest & Othr Incm                 130     -5.8%         127      2.4%         257       150      18.1%    407            156      25.8%        563        9.7%

                                                                                                                                                        (continued)
Figure 1.4 (continued)


Cisco Systems
Income Statement         1Q05      Y/Y %     2Q05      Y/Y %      1H05      3Q05      Y/Y %   9mos05     4Q05      Y/Y %     2005      Y/Y %
Mnrty Intrsts/Othr            40      0.0%        17     -87.5%        57         -                 57         -                  57    -69.5%
    Pretax Income          1,958     28.1%     1,957       8.3%     3,915     1,968     14.4%    5,883     2,153     11.2%     8,036     14.9%
    PF Pretax Income       2,018               2,058                4,129     2,078              6,207     2,259               8,466
Incm Taxes                   562     27.1%       557      7.9%      1,119       563     10.4%    1,682       613     10.3%     2,295     13.4%
PF Incm Taxes                565                 576                1,141       582              1,723       633               2,356
Tax Rate                  28.7%      -0.8%    28.5%      -0.3%     28.6%     28.6%      -3.5%   28.6%     28.5%      -0.9%    28.6%      -1.3%
PF Tax Rate               28.0%               28.0%                27.6%     28.0%              27.8%     28.0%               27.8%
Accnt Chges
Net Income                 1,396     28.5%     1,400     93.4%     2,796      1,405    16.0%     4,201    1,540      11.6%    5,741      15.6%
Pro Forma Net Oprtns       1,453     22.8%     1,482     12.7%     2,988      1,496    10.1%     4,484    1,626       9.9%     6,110     12.7%
Basic Shares               6,635     -4.3%     6,521      -5.1%    6,578      6,435    -5.6%     6,530    6,366      -5.5%    6,489       -5.1%
Diluted Shares             6,773     -4.7%     6,652     -6.4%     6,713      6,541     -7.5%    6,655    6,480      -6.6%    6,612      -6.3%
Rptd Basic EPS              0.21     34.3%      0.21    103.8%      0.43       0.22    22.9%      0.64     0.24      18.1%     0.88      21.8%
Rptd Dltd EPS               0.21     34.9%      0.21    106.7%      0.42       0.21    25.5%      0.63     0.24      19.4%     0.87      23.4%
Adjstd Basic EPS            0.22     28.3%      0.23     18.8%      0.45       0.23    16.6%      0.68     0.26      16.3%     0.93      19.6%
Adjstd Dltd EPS             0.21     28.9%      0.22     20.4%      0.44       0.23    19.1%      0.67     0.25      17.6%     0.92      21.2%
                           Phase 1: Income Statement and Margin Model, Part 1   •   23



inertia, analysts covering such “start-ups” as Cisco and Microsoft still routinely
exclude noncash charges. Once Financial Accounting Standards Board (FASB)
established FAS 123R, it was no stretch to lard stock option compensation in with
the other noncash items.
      The exclusion of nonrecurring items is more defensible—or would be, if
certain companies did not always manage to find something that they can sneak
across the line as “nonrecurring.” Companies whose balance sheets are as large
as those of small nations will routinely be incurring nonrecurring costs related
to new deals and to cleaning up the mess from past failed deals.
      Even well-structured acquisitions executed at fair prices can now trigger mas-
sive impairments once the market sours. Nearly every significant asset purchase
entails recognition of goodwill on the balance sheet; in our experience, this will
equate to anywhere from 40% to 70% (or more, rarely less) of the purchase price. In
our view, the FASB doomed investors to a regular barrage of nonrecurring charges in
the form of goodwill impairment when it changed the treatment of goodwill.
      Goodwill used to be depreciated, meaning that the excess from assets
acquired at a premium in good times would be steadily worked down on a sched-
uled basis. True, as in any game of musical chairs, some companies closed major
asset deals immediately before a severe market downturn, prompting significant
write-downs. But in the aggregate, depreciation of goodwill gradually cleansed
balance sheets of excess goodwill and prevented immense write-downs of a very
vulnerable asset class in tough times.
      Beginning in mid-2001, under FASB standard 142, goodwill was no longer
depreciated. Instead, it was to be left on the balance sheet and impaired as need
be to reflect changes in peer-group values. Now that the balance sheet–cleansing
mechanism of depreciation is gone, every down cycle triggers a huge number of
huge goodwill impairments. Goodwill write-downs and other impairments flow
through the income statement and onto retained earnings on the balance sheet,
thus impacting stockholders’ equity; that is playing havoc with return on equity
and debt-to-capitalization ratios, in our view. Figures 1.5 and 1.6 show the con-
sequences of this change in accounting treatment, in the form of massive good-
will impairments at Vishay Intertechnology and ADC Telecom taken during the
market downturn of 2007–2009.
      The upshot for analysts is that every income statement model should be
built to accommodate the high likelihood of a major impairment event—and the
concomitant high likelihood that investors, for an indeterminate period, will
value the company on adjusted earnings rather than GAAP earnings. In the age
of goodwill impairment as opposed to goodwill depreciation, every income state-
ment model must be ready to accommodate pro forma earnings. Think of it as a
Figure 1.5

Vishay, a company with a healthy and disciplined business development strategy, impaired goodwill in every quarter of 2008 as the market collapse drove down
asset values.

Vishay Intertech
Income Statement               1Q08                 2Q08                   1H08        3Q08             9mos08       4Q08                   2008
 Net Sales                      733,313      11%     774,364       8%     1,507,677    739,092       1% 2,246,769    575,442      -21%     2,822,211         -0%
 COGS                          560,850       16%    594,645       11%     1,155,495     579,591      4% 1,735,086     484,134     -14%     2,219,220          4%
 Loss Purchs Cmmtmnts                                                              -                             -      6,024                  6,024
 Gross Profit                   172,463       -2%      179,719       1%      352,182    159,501      -9%    511,683     85,284     -49%      596,967         -14%
 SG&A Costs                    119,063       11%      121,021       7%     240,084     112,844       3%   352,928      97,951     -11%       450,879          3%
 Restructuring                  18,202      798%        8,909     618%        27,111     6,849     -31%     33,960     28,577                 62,537
 LT Prchs Cmts/Trmndt Tender                                                             4,000                            878                    878
 Amortzn                                                                           -                              -                                -
 Wrtdwns/Gdwll Imprmnt            4,195              800,000                804,195   357,917             1,162,112  565,257              1,727,369
 Purchased R&D                                                                     -                              -                                -
     Operating Costs           141,460               929,930              1,071,390    481,610            386,888 692,663                  1,079,551
     PF Oprtng Costs           116,682                119,811               236,493   108,330             344,823      97,951                442,774
 Operating Income               31,003      -53%    (750,211)   -1332%    (719,208) (322,109)     -686%    124,795 (607,379)    -1803%    (482,584)        -301%
 PF Oprtng Incm                 55,781                59,908                115,689      51,171            166,860    (6,643)                160,217
 Interest Expense                6,584       -8%        6,078     -18%        12,662     4,873     -35%      17,535     6,729      2%         24,264        -15%
 Trmt IR Swp/Xtngsh Debt                                                           -    13,601               13,601                           13,601
 Extrdny Gain (Loss)                                                               -                              -                                -           0%
 Other                             198      -103%     (4,673)       11%      (4,475)   (6,853)     318%   (11,328)    (3,548)       -6%     (14,876)         -11%
 Minority Interest                 478        65%         269        4%          747       144     -67%         891       173      -12%        1,064         -10%
 Ttl Othr Expnss                 7,260     1738%        1,674      -52%        8,934    11,765      89%     20,699      3,354       10%       24,053          83%
 Pretax Income                  23,743       -64%   (751,885)   -1409%    (728,142) (333,874)     -786%    104,096 (610,733)    -1973%    (506,637)        -323%
   ProForm Prtx Incm            48,521                 58,234               106,755    39,406               146,161   (9,997)                136,164
                                                                                                                                                       (continued)
     Figure 1.5 (continued)


     Vishay Intertech
     Income Statement            1Q08                 2Q08                   1H08        3Q08               9mos08         4Q08                    2008
      Income Taxes                   6,173    -61%    (10,194)      -166%     (4,021)    (21,007)     -281%  (25,028)         36,215      70%         11,187     -83%
        ProForm Incm Txs            12,616              15,141                27,756        5,517              33,273          2,599                 35,872
      Tax Rate                     26.0%       8%        1.4%        -95%       0.6%        6.3%       -74%   -24.0%          -5.9%     -109%         -2.2%     -108%
      PF Tax Rate                  26.0%                26.0%                  26.0%       14.0%               22.8%         -26.0%                  26.3%
      Discontinued Operations   42,136.00                                                                                (5,690.00)              (5,690.00)
      NetIncome                  (24,566)    -149%    (741,691)    -1920%    (724,121)   (312,867)    -989%   129,124     (652,638)    -13479%    (523,514)     -441%
      PF Net Income                35,906                43,093                78,999       33,889            112,888      (12,596)                 100,292
      Basic Shares Out            186,343        1%     186,371         1%    186,357      186,651       0%   186,455       186,544         0%      186,477         1%
      Diluted Shares Out          186,540      -13%    186,820         -3%    186,680      187,100      -3%   186,820       186,544        -6%      186,751        -6%
      Rprtd Basic EPS               (0.13)   -149%        (3.98)    -1911%      (4.11)       (1.68)   -988%     (5.79)        (3.50)   -13415%        (9.29)    -1416%
      Rprtd Diluted EPS             (0.13)   -152%        (3.97)   -1892%       (4.10)       (1.67)   -993%     (5.77)        (3.50)   -12071%        (9.29)   -1448%
      Adjstd Basic EPS                0.19    -32%          0.23      -10%        0.42         0.18    -29%       0.61        (0.07)     -134%          0.54      -46%
      Adjstd Diluted EPS              0.19    -27%          0.23      -10%        0.42         0.18    -28%       0.60        (0.07)     -135%          0.54      -44%
25
Figure 1.6

ADC Telecom took a huge impairment charge in the first quarter of its September 2009 fiscal year, resulting in a deep GAAP loss for the year.

ADC Telecom
Income Statement            1Q09     Yr/Yr %       2Q09     Yr/Yr %      1H09        3Q09     Yr/Yr % 9mos09          4Q09E Yr/Yr %             2009E     Yr/Yr %
Prodct Sales                  218.50      -27%      239.30      -32%        457.8      245.8      -28%    703.6         152.54  -50%                856.1     -34%
Service Revenue                35.80       15%       35.80      -10%         71.6       37.6        -8%   109.2          22.50  -50%                131.7     -16%
     Net Sales                 254.3     -23%        275.1      -30%        529.4      283.4      -26%    812.8          175.0  -50%                987.8     -32%
Product COGS                  147.50                155.40      -31%        302.9      155.1       -31%   458.1          96.30                    554.36      -34%
Service COGS                   28.10                 29.60      -14%         57.7       29.6       -15%    87.2          18.34                    105.59      -19%
     Net COGS                  175.6      -19%       185.0      -29%        360.6      184.7      -29%   545.3           114.6  -50%               659.9      -32%
 Gross Profit                    78.7      -30%         90.1     -32%        168.8       98.7      -20%    267.5           60.4  -50%                327.9     -33%
 Research & Development         19.0       -3%         18.4     -16%         37.4        17.3     -20%     54.7           10.1  -51%                 64.8     -22%
 Selling & Administration       71.9       -3%         66.1     -15%        138.0       62.9       -16%  200.9            37.6  -63%               238.5      -27%
 Intngbls Amrtzn                   -                      -                      -          -                 -               -                         -       0%
 In-Prcss R&D/Rtntn                -                      -                      -          -                 -               -                         -
 Impairment Charge             413.5        0%          0.7       0%        414.2         0.1        0%   414.3               -   0%                414.3       0%
 Restructuring Charge            0.5     -58%           7.3       0%           7.8        5.3    -763%     13.1             5.0 -77%                 18.1      19%
 Operating Costs & Expns       504.9     389%          92.5     -15%        597.4       85.6       -18%  683.0            52.7  -52%               735.7       72%
 PF Optrng Costs & Expns        82.6                   78.3     -21%        160.9       72.0              232.9           42.7                     275.6
    Operating Earnings       (426.2)   -4889%         (2.4)    -110%      (428.6)       13.1      -28%  (415.5)             7.7 -28%              (407.8)    -761%
    PF Oprtng Ernngs             6.1     -68%          11.8     -65%          17.9      26.7         2%    44.6            17.7 -62%                 62.3     -50%
 Net Interest Income               -                      -                      -          -                 -               -                         -       0%
 Gain (lss) Investments                               (0.7)
 Gain on Patent Settlment
 Other Loss (Income)              20.3    -59%           5.8     -64%         26.1        7.1      -18%        33.2         6.0    -2100%           39.2      -61%
 Pretax Income                (446.5)    1140%         (7.5)    -194%      (454.7)        6.0      -31%     (448.7)         1.7      -88%        (447.0)    1070%
 PF Pretax Income               (11.4)                   6.7                 (4.7)       19.6                  14.9        11.7                     26.6
 Income Taxes                    (4.0)   -367%           1.4     -26%        (2.6)        0.4      -86%       (2.2)         0.2      -88%          (2.0)    -133%
     Figure 1.6 (continued)


     ADC Telecom
     Income Statement            1Q09      Yr/Yr %    2Q09     Yr/Yr %    1H09       3Q09    Yr/Yr % 9mos09        4Q09E Yr/Yr %    2009E    Yr/Yr %
      PF Income Taxes               (4.0)                  1.4               (2.6)       4.3               1.8         2.6               4.3
      Tax Rate                        1%                -19%                  1%         7%                0%         10%                0%
      PF Tax Rate                   35%                  22%                 54%        22%               12%         22%               16%
      Income Cntng Oprtns         (442.5)     1080%     (8.9)     -246%    (452.1)       5.6      -3%  (446.5)         1.5   -88%    (445.0)    902%
      PF Incm Cntng Oprtns           (7.4)                 5.3               (2.1)      15.3              13.1         9.1              22.3
      Incm/Lss Dscntd Optrns        (0.3)                (1.3)               (1.6)     (6.4)             (8.0)           -             (8.0)
      Loss/Gain Sale
      Net Income                  (442.8)    1103%      (10.2)    -273%    (453.7)      (0.8)   -112%    (454.5)       1.5   -88%    (453.0)   981%
      PF Net Income                  (7.4)                  5.3              (2.1)       15.3               13.1       9.1              22.3
      Basic Shares Outstndng         99.4     -15%        96.6     -18%       98.0       96.6    -18%       97.5      96.7   -18%       97.3    -17%
      Diluted Shares Outstndng       99.4     -33%        96.9     -18%       98.2       97.8    -17%       98.0      97.9   -17%       98.0    -16%
      Cntng Ops EPS Basic          (4.45)               (0.09)              (4.61)       0.06             (4.58)      0.02            (4.57)
      Cntng Ops EPS Diluted        (4.45)               (0.09)              (4.61)       0.06             (4.55)      0.02            (4.54)
      Reprtd EPS Basic             (4.45)    1324%       (0.11)   -311%     (4.63)     (0.01)   -106%     (4.66)      0.02   -85%     (4.65)   1201%
      Reprtd EPS Diluted           (4.45)    1986%       (0.11)   -226%     (4.62)     (0.01)   -106%     (4.64)      0.02   -89%     (4.62)   1192%
      Dscntd Ops EPS Basic         (0.00)     -151%     (0.01)     692%     (0.02)     (0.07)   -809%     (0.08)         -     0%     (0.08)
      Dscntd Ops EPS Diluted       (0.00)     -164%     (0.01)     693%     (0.02)     (0.07)   -804%     (0.08)         -     0%     (0.08)
      Adjstd EPS Basic             (0.07)     -124%       0.05     -85%     (0.02)       0.16     -42%      0.14      0.09   -49%       0.23    -79%
      Adjstd EPS Diluted           (0.07)     -126%       0.05     -86%     (0.02)       0.16    -42%       0.14      0.09   -49%       0.23    -80%
27
28   •   Income Statement Presentation



form of disaster planning. You don’t try to buy flood insurance when your desk
is floating out of the room. Similarly, we can’t try to adjust on the fly for the cata-
strophic impairment, particularly given that well-intentioned accounting stan-
dards (by definition, the worst kind) have nearly preordained such an event as a
regular or at least cyclical occurrence.
       A final thought on goodwill impairment and its consequences, and this
counts as my first real rant. In some ways, the impairment of goodwill has accus-
tomed investors to accept non-GAAP, or pro forma, results as readily as GAAP
results. In a kind of cascading effect, that may have led aggressive managements
in areas outside traditional growth niches to begin calling out noncash and non-
recurring events that were previously treated as ordinary costs of doing business.
I’m suggesting that by opening the floodgates of goodwill impairment, FASB may
have inadvertently encouraged a lot of questionable callouts to wash over the
reporting process.


Amending Historical Income Statement Data

Now, an important point: analyst convention, based on collusion between the
consensus-gathering companies and the analysts, will determine (1) whether
GAAP or non-GAAP earnings drive the investment decision for a particular
company, and (2) if non-GAAP prevails, which line items are regularly excluded.
Believe me when I say there is much variation on this matter among the roughly
5,000 stocks trading on the NYSE and the Nasdaq Composite Index. A call to the
consensus manager (e.g., Bloomberg, Thomson-First Call) can sometimes pro-
vide an answer; paying close attention to the footnotes in a company’s results
release will also provide some clues.
      So let’s retrace our walk down the income statement, ready to add a few new
elements. Remember income statement convention will vary from CFO to CFO.
Some of our additions will be to “normalize” the income statement for easy cross-
industry comparisons; others will be to provide a ready means for easy pro forma
calculation. Key point: the ability to call out and isolate items to create pro forma
results does not mean we say to do so; we merely want to have that option as the
need arises. Note, we use the term pro forma for non-GAAP or adjusted line items
at least partly because the acronym PF fits easily on the amended-item line.


Pro Forma Line Items
Our first task is to adjust the income statement so we can record, measure, and
analyze GAAP and pro forma results—and modeled results—simultaneously.
                           Phase 1: Income Statement and Margin Model, Part 1   •   29



The alternative is to maintain separate income GAAP and pro forma income
statements. That is a repetitive task, particularly if one-time events really are
scarce.
      As an overview, we will generally insert lines to sum pro forma operating
costs and to show the following:

     •   Pro forma operating earnings
     •   Pro forma pretax income
     •   Pro forma income tax
     •   Pro forma tax rate
     •   Pro forma net income
     •   Pro forma basic and diluted EPS

      For companies with meaningful joint venture exposure, we will also show PF
net income before MI and EE and then reflect the equity income and minority
interest inputs to arrive at pro forma net income. We generally show basic and
diluted share count only once, on a GAAP basis. For companies with significant
convertible debt, or for companies with a chronically huge gap between GAAP and
pro forma results, we might include a line item showing PF diluted share base.
      For the most part, GAAP and pro forma revenues are not materially differ-
ent from each other. Non-GAAP cost of goods sold (COGS) can be reduced,
typically by stock option compensation. You have the discretion to create a new
line item for non-GAAP COGS. For some larger companies in which this is a
large item, we will include a line item for non-GAAP COGS. For the smaller firm,
we are less apt to create this line item.
       Some companies, in their income statement presentations, display gross
profit, defined as the difference between revenues and COGS; some don’t. We
always will; so if the presentation does not include a line for gross profit (revenue
minus COGS), insert a line and display gross profit. Again, we do not typically
display PF gross profit.
      Of the main operating cost items, SG&A and R&D are typically big sites for
stock option compensation. Some nonrecurring or irregular items, such as legal
costs, can sometimes be carried within these lines. Below these main operating
cost line items we will list noncash and nonrecurring items; we will call out each
separately so they can be easily excluded from non-GAAP computations.


Making Room for Nonrecurring and Noncash Items
In Chapter 11, we will discuss companies’ increasing propensity for behavior that
seems destined to trigger “nonrecurring” events. Putting aside the occasional
30   •   Income Statement Presentation



positive one-time event, such as a legal settlement resulting in a gain or a tax
reversal, most such one-time events involve sizable impairments to assets or to
goodwill. We will suggest in fact that the system as currently structured seems to
make such one-time and nonrecurring events commonplace, thanks to ill-timed
and ill-conceived acquisitions, sudden veers in the economic cycle, and FASB
standard 142, which compels goodwill tests and impairment rather than goodwill
depreciation.
      For now it is sufficient to know that these events are more frequent than in
the past and that our models must be able to accommodate them. So as we amend
our income statement presentation, we will include at least two lines to accom-
modate these errors when acknowledged. We’ll call one line item restructuring
and the other impairment.
      In practice, I make use of at least three lines, the third usually being called
acquisition-related. There’s a certain crazy and circular logic to this presentation.
Without fail, companies overpay for acquisitions, eventually take restructuring
costs for their hubris, and impair oodles of goodwill generated by ill-advised
acquisitions.
      In our adjusted presentation, we will always make room for a final operat-
ing cost item that we call “other.” Many companies already have such a line item.
However, even if a company does not list an “other” category, we’ll include one
anyway. In short, if we’ve modeled multiple forward quarters and years and sud-
denly at the company’s behest need to sandwich in another line item—a very
common occurrence—we would need to change all our sum-of–operating cost
formulas. We would have to, that is, if we were unprepared. If a company hatches
such a line item—and I’ve seen some doozies—we’ll insert a line directly above
“other” to accommodate the new curiosity. This way, we don’t need to amend the
summary formulas that are in every quarterly calculation of GAAP and pro
forma operating costs.
      Within the actual income presentation from the company, conventions will
vary, and companies may or not sum operating costs. We always will; as need be,
insert a line for GAAP operating costs. Insert a line and call it pro forma operat-
ing costs. So now we have two operating cost tallies, one for GAAP operating
costs and one for pro forma operating costs.
      GAAP operating cost sums the headline operating costs—SG&A and R&D,
or alternately S&M, R&D, and G&A—along with the called-out and thus easily
excluded noncash and nonrecurring items. Non-GAAP or pro forma operating
income sums only the headline items and is adjusted as needed to exclude items
not called out.
                           Phase 1: Income Statement and Margin Model, Part 1   •   31



Adjusting Historical Quarters for Stock Option Compensation
Most nonrecurring and noncash items earn their own line in the income state-
ment and are easily excluded. Remember, FAS 123R stock option compensation is
not an income statement line item. It can sometimes be tricky to find and thus
sometimes difficult to exclude from historical comparisons.
       For our historical determination of pro forma operating costs, particu-
larly if we are going to regularly exclude stock option compensation, we need
to find a historical basis. Every public company in the United States reports its
quarterly results. If covering analysts are accustomed to excluding noncash
items, companies usually cooperate by listing them. In particular, we’ll typi-
cally be able to find allocation of stock option compensation per quarter div-
vied up among the recipients by line item (i.e., by COGS, SG&A, and R&D).
But your best source is not the 10-Q, which by convention still tends to bury
this data. Consult the quarterly results release issued by the company and
archived on its Web site; here you’ll typically find the stock option compensa-
tion detail by line item.
       Most of the hourly workers toil in the factories, while most of the salaried
workers are in sales, engineering, and other white collar occupations. Stock
option compensation is usually minor as a proportion of COGS (typically less
than 1%), while it can be meaningful in the main operating cost lines (3% to
10%, or more, of operating costs).
       There are several ways to capture or re-create stock option compensation in
historical presentations with the goal of excluding FAS 123R from PF operating
costs. One, we can measure the percentage of individual line-item costs repre-
sented by FAS 123R for several quarters and adjust each line item by the average
of several periods. Thus, if we determine that, on average, FAS 123R amounts to
7% of SG&A and 9% of R&D, in our PF operating cost calculation we will refer-
ence the actual Excel cells representing SG&A and R&D (let’s say A7 and A8,
respectively). For pro forma SG&A, we will use the formula A7*.93; and for
R&D, A8*.91.
       The second way would be to capture the historical data, culled from the pro
forma portion of the press release, and list the actual FAS 123R contributions by
line item in a column immediately below that quarter’s income statement. Thus,
if we learn that GAAP COGS contains $8 million in FAS 123R costs, we list it,
along with the $20 million included in SG&A and the $23 million included in
R&D; hypothetically, these are in cells A40, A41, and A42. In our pro forma
operating cost calculation we will reference the actual Excel cells representing
32   •   Income Statement Presentation



SG&A and R&D (again, A7 and A8, respectively) and adjust as follows: A7 A41
and A8 A42. Notice we did not reference cell A40? We’ll get to it.
      Another option is to calculate FAS 123R stock option compensation cost as
a percentage of GAAP operating expenses. Remember, stock option compensa-
tion for the middle tier of engineers and salespeople is not a given (as it too often
is in the executive suite) but is based on performance. When GAAP operating
costs are rising, the cause would be sloppy execution (seldom) or, more often,
rising sales performance and thus rising compensation. We put four to eight
quarters of historical P&L in our model for just such situations where we need to
see the precedent.
      To proceed, measure FAS 123R stock option compensation as a percentage
of GAAP operating costs for four to eight historical quarters; determine any aver-
age; check for any seasonal outliers; and use that seasonally adjusted percentage
in appropriate quarters. Asset-light companies with a higher proportion of vari-
able costs [e.g., software firms or electronics original equipment manufacturers
(OEMs) that rely on contract manufacturers] tend to have a higher percentage
figure, often 5% or more. Traditional asset-intensive companies with high fixed
costs will have a lower percentage, say 2% to 3%, in FAS 123R compensation as
a percentage of operating costs.
      In Figure 1.7, in shaded cells at the bottom of the worksheet we see Juniper’s
historical stock option compensation costs by line item; these items are also
expressed in percentage terms, to give some sense of their size and scope. These
can easily be deducted from PF operating costs, along with more visible nonre-
curring and noncash line items.


Amending Historical Adjusted Operating and Pretax Income
Beneath GAAP operating cost and pro forma operating cost, we want to have a
line for GAAP operating income and a line for pro forma operating income.
GAAP operating income consists of gross profit minus GAAP operating cost. Pro
forma operating profit consists of adjusted gross profit minus pro forma operat-
ing cost.
      How do we adjust gross profit so it is reflected in our pro forma operating
cost? Again, we can go one of two ways. Let’s begin with the premise that GAAP
gross profit is in cell A4. The first method by which we model adjusted operating
profit is to increase gross profit in this formula to reflect stock option compensa-
tion cost as a percentage. So in the hypothetical cell A13, which shows pro forma
operating profit, we use the formula (A4*1.005) A10. The second method by
which we model adjusted operating profit is to increase gross profit in this for-
Figure 1.7

In 2006, Juniper’s FAS 123R stock option compensation typically amounted to 6% to 8% of GAAP operating costs. (The 2Q06 outlier reflects the $1.2 billion impair-
ment in that quarter.) Note that less than 10% of stock option compensation goes to production workers (expressed in COGS), while over 90% goes to the salaried
staff (represented in operating cost line items such as R&D, S&M, and G&A).


Juniper Networks
Income Statement                1Q06       Y/Y %      2Q06       Y/Y %       1H06          3Q06        Y/Y % 9mos06           4Q06       Y/Y %     2006     Y/Y %
 Product Revenue                 474,125       21%    467,237        10%        941,362    467,524          0% 1,408,886      483,500        -1% 1,892,386       7%
 Service Revenue                  92,589      63%      106,330      53%         198,919     107,090        34% 306,009         112,330       29%    418,339     43%
 Revenue                         566,714      26%     573,567        16%      1,140,281     574,614         5% 1,714,895      595,830         4% 2,310,725      12%
 Prdct COGS                     140,995                144,843                 285,838      147,906                433,744    133,866               567,610
 Srvc COGS                        43,952                47,849                    91,801      41,717                133,517     56,379             189,896
 Cost of revenue                 184,947      29%      192,692      24%        377,639      189,623        11%     567,261     196,624        8%   757,506      16%
 Gross profit                     381,767      25%     380,875       13%        762,642     384,991          3% 1,147,633      405,595         3% 1,553,218      10%
   Research and development      113,688      49%      120,449      48%         234,137    123,542         37%     357,679     128,103       29%   485,782      40%
   Sales & Marketing             129,429      42%      141,958      39%         271,387     143,653        24%     415,040     148,958       18% 563,998        29%
   General and administrative     23,099      49%       25,811      68%           48,910     25,858        -4%       74,767     26,812       65%    101,579     37%
   Amrtzn & Deferred Comp         23,221       6%       22,000      -8%           45,221     22,000       -25%       67,221     22,000       -7%     89,221    -10%
   Restructuring                   1,404                                           1,404                              1,404          -                1,404
   Impairmnt/Charitabl                               1,283,421               1,283,421                           1,283,421                       1,283,421
 Total expenses                 290,841        42% 1,593,638       630%      1,884,479     315,053         18% 2,199,532      325,873        18% 2,525,406     161%
 PF Total expenses              245,031                267,009                  512,040    272,299                 784,339    286,895             1,071,235
     Operating income (loss)     90,926       -10% (1,212,763)    -1114%    (1,121,837)      69,938       -36% (1,051,899)      79,722      -32% (972,187)    -319%
     Pro Forma Op Income        139,985                126,702                 266,687      121,661                388,348     120,105             508,453
 Interest, othr incm (expns)      20,767       95%      23,882      78%          44,649      26,143        68%       70,792     26,143       37%     96,935     65%
 Intrst Expense                  (1,089)                 (750)                   (1,839)      (750)                 (2,589)      (750)              (3,339)
 Gain Debt Extgt/Invstmnts                                                             -                                  -                               -
     Pretax Income              110,604       -1% (1,189,631)     -1002%   (1,079,027)      95,331        -24% (983,696)       105,115      -22% (878,591)   -275%
     PF Pretax Income           159,663               149,834                  309,497     147,054                 456,551    145,498              602,049
                                                                                                                                                         (continued)
Figure 1.7 (continued)


Juniper Networks
Income Statement           1Q06        Y/Y %    2Q06         Y/Y %      1H06         3Q06        Y/Y % 9mos06         4Q06        Y/Y %    2006       Y/Y %
 Taxes                       34,841        -3% (344,993)       -904%    (310,152)      27,646       -33% (282,506)      25,228       -11% (257,278)     -274%
 PF Taxes                    46,302               43,452                   89,754      42,646              132,400      34,920               167,319
Tax rate                       32%                  29%                      29%         29%                  29%         24%                   29%
PF Tax rate                    29%                  29%                      29%         29%                  29%         24%                   28%
 Net income (loss)           75,763         0% (844,638)      -1049%   (768,875)       67,685       -19% (701,190)      79,887       -25% (621,313)     -276%
 Pro Forma Net Incm         113,361              106,382                  219,743      106,119             324,151     112,636              438,529
 Shares Basic & diluted    565,927          4%   565,927         4%      565,927     563,097          0% 564,984      560,282         -1% 563,808          2%
 Pro Forma Shares          603,589          3% 603,589           2%      603,589      600,571        -1% 602,583      597,568         -2%   601,329        1%
 Rprtd Basic EPS                0.13               (1.49)                   (1.36)        0.12               (1.24)        0.14                (1.10)
 Rprtd Dltd EPS                 0.12       -7%     (1.40)     -1029%        (1.27)        0.11      -19%     (1.16)        0.13      -24%      (1.10)  -287%
 Pro Forma Basic EPS           0.20                  0.19                     0.39        0.19                 0.58       0.20                  0.78
 Pro Forma Dltd EPS             0.19      20%        0.18       -0%           0.36        0.18       -7%       0.54        0.19       -4%       0.73       1%

 Cash & Equivalents        2,614,304             2,614,304                           2,614,304                        2,614,304
 Debt                       399,944               399,944                             399,944                          399,944

 COGS                          1,883                 1,968                               2,101                           1,571
 R&D                          10,013                 9,407                              9,364                            7,000
 S&M                           7,627                8,486                               8,071                             7,121
 G&S                           3,545                 3,315                              3,319                            2,857
   FAS 123R                  23,068                 23,176                             22,855                           18,549

 % GAAP Operating Costs        7.9%                  1.5%                                7.3%                             5.7%

 FAS 123R % Mnfctng            8.2%                  8.5%                               9.2%                              8.5%
 FAS 123R % R&D and SG&A      91.8%                 91.5%                              90.8%                             91.5%
                           Phase 1: Income Statement and Margin Model, Part 1   •   35



mula to reflect stock option compensation cost as a modeled number. So in the
hypothetical cell A13, which shows pro forma operating profit, we use the for-
mula (A4 A4) A10.
      Now we have GAAP operating income and PF operating income. Next are
the nonoperating items, principally interest income and interest cost. Do these
vary for GAAP and PR purposes? Not typically, though some companies (usually
very large or very global ones) may call out a slight difference. The catch-all cat-
egory, “other,” however, may be excluded by analysts and/or the company from
pro forma results.
      GAAP operating income less net interest cost and less nonoperating “other”
is equal to pretax income. Pro forma operating income less (possibly adjusted)
net interest cost and (possibly excluded) nonoperating “other” is equal to pro
forma pretax income. We therefore need to create a line to accommodate PF
pretax income. It follows that if we have two calculations of pretax income, we
need to tax them separately. Accordingly, we create a separate line for GAAP taxes
and PF taxes.


Finishing the Historical Adjustments
The standard corporate tax rate in this country is 34%. But most midsized and
larger public companies, and not a few small-capitalization companies, have at
least some foreign pretax income, which is taxed at different rates around the
world.
      Honestly achieved losses (e.g., through simple incompetence, badly mis-
judged markets, and wildly overpriced asset buys) trigger losses that are taxed at
about the corporate rate. So, if you honestly lose $1 million, and assuming a 34%
tax rate, your tax bill might be a credit in the amount of about $340,000. Now
that goodwill is no longer depreciated but impaired, huge impairments flowing
through the income statements are a commonplace. Yet impairment-related
losses rarely trigger commensurate tax breaks; and again, the issue is origination
of loss. If a company impairs assets, including goodwill, in a mature market but
makes money in other markets, that $1 million loss may actually be accompanied
by GAAP taxes, not tax losses. More significantly, impairments can also result in
significant deferred tax valuation allowance charges.
      Given the wide variety of special items, regional markets, and other effects,
I find it particularly challenging to model GAAP taxes. By contrast, modeling PF
taxes are a breeze. What’s interesting here, and speaking anecdotally, is that PF
income taxes tend to be (somewhat) predictable, while GAAP taxes vary enor-
mously. Having adopted the fiction of PF earnings—a variation on that other
36   •   Income Statement Presentation



collective hallucination known as the GAAP income statement (but now I’m
starting to rant)—CFOs add some “stability” to the picture by forecasting a PF
tax rate. Since impairments don’t figure in PF results, PF taxes attempt to simu-
late a rate appropriate to those “honestly achieved losses” mentioned previously.
Surprisingly often, PF taxes are delivered at the company-projected rate of 28%,
34%, or whatever.
       Below the GAAP and PF income taxes, we like to include the tax rate rep-
resented in percentage form as a percentage of pretax income. We italicize these
lines, which represent the only percentage figures shown within the individual
quarter P&L columns. True, these percentages are represented in the margin
analysis immediately below. But again, within our income statement presenta-
tion, we always emphasize visual representation of useful information.
       Earlier, we briefly discussed the significant below-the-line items, including
minority interest, equity income, accounting adjustments, and discontinued
operations. As a guideline rather than a rule, PF net income needs to accommo-
date minority interest and equity income but only rarely accounting adjustments
and discontinued operations. GAAP net income needs to accommodate all these
items.
       Net income is the difference between pretax income and GAAP taxes and
all items, whereas pro forma net income is the difference between PR pretax
income and PF taxes and (most typically) equity income and minority income.
Below GAAP and PF net income, and regardless of their positioning within the
company presentation, we position basic and diluted shares. We do not typically
create a line item for PF diluted shares outstanding, except for those companies
with significant convertible debt or huge chronic differences between GAAP and
PF results. Your signal that a company has lots of convertible debt, without look-
ing at the balance sheet (shame on you), is a large gulf between diluted shares
outstanding and basic shares outstanding.
       If a company has no discontinued operations to speak of, the company’s
income statement presentation will not show this line item. Nevertheless, every
company at some point will hold an asset for disposition. Given the eternal dis-
connect between sellers and buyers, this could be a long process, one that will
stretch across several (from the CFO’s perspective) nail-biting quarters.
       Accordingly, below basic and diluted shares, we allocate six lines: basic and
diluted reported EPS; basic and diluted discontinued operations EPS; and basic
and diluted pro forma EPS. We’re almost at the bottom of our remodeled income
statement; but there are a couple of must-haves for the conscientious modeler. We
already mentioned that for companies whose valuations are driven by pro forma
                          Phase 1: Income Statement and Margin Model, Part 1   •   37



earnings, and because the analyst convention is to exclude stock option compen-
sation, we have the option of listing the FAAS 123R numbers for each line item.
      Casual investors may be surprised that many dividend-paying companies
do not include their dividends within the income statement presentation. That is
actually appropriate; the income statement is a record of operations, while the
dividend decision is a reflection of financial policy. Many companies do include
the dividend at the bottom of their income statement presentation, usually
because they’re feeling pretty good about the payout.
      Within our income statement presentation, we’re always going to situate the
quarterly dividend beneath pro forma diluted EPS, not because we’re feeling par-
ticularly good about anything but because—linked to another worksheet—it will
prove useful later on.
      We’ve now built the structure for modeling the income statement. As noted,
we’ve made several preemptive adjustments based on the high likelihood that
corporate or economic events will fracture the relationship between GAAP results
and adjusted results. Now we need to model the line-item inputs that provide the
basis for our forecast GAAP and forecast adjusted results.
This page intentionally left blank
                                                   Chapter 2
           PHASE 1: INCOME
           STATEMENT AND MARGIN
           MODEL, PART 2




Margin Analysis
We’ve now amended the historical income statement presentation to make it more
user-friendly. We need a visual representation of the value of these inputs before we
can begin modeling individual line items for forward quarters and interim and
annual periods. We’re going to conduct all of the following discussion under the
heading “Margin Analysis.” By that we generally mean representing income state-
ment items as a percentage of revenue. But there will be some exceptions.
      As always, what the company gives us will partly inform our margin model.
Many companies include as part of their income statement presentation a break-
out of product and service revenues. If so, the first lines in our margin analysis
represent percentage of revenues for product and percentage of revenues for ser-
vices. If revenue is presented as a single number, it earns no place in our margin
analysis.
      We next represent cost of goods sold (COGS) as a percentage of sales; for
this and all future lines, we’ll express percentages as “(Item) % Revenue.” Below
COGS as a percentage of revenue, we have gross profit, which is simply revenue
minus COGS and represented as a percentage of revenue. In the next few lines,
we represent R&D and, depending on presentation, either S&M and G&A, or
alternately SG&A, as a percentage of revenue.
                                                                                  39
40   •   Income Statement Presentation



      Beyond these major cost categories, it makes sense to represent amortiza-
tion of intangibles. The amortization schedule typically though not always lends
steadiness and predictability to this account; as such, it most often impacts GAAP
net income in a predictable way. The item being amortized—acquired intangi-
bles—resides not on the income statement but on the balance sheet. In our mar-
gin analysis bloc, we will represent amortization as a percentage of revenue. But
in our modeling of future quarters, whenever possible we will be modeling intan-
gibles amortization as a percentage of the balance sheet item “acquired
intangibles.”
      The other noncash and nonrecurring items impacting operating costs will
also be portrayed as a percentage of revenue. We then have a line for GAAP oper-
ating costs as a percentage of revenue, and PF operating costs as a percentage of
revenue. Below this are lines for GAAP operating income as a percentage of rev-
enue and PF operating income as a percentage of revenue.
      Beneath these are the nonoperating lines for items that impact pretax
income; depending on presentation, these will include interest cost, interest
income, and/or net interest expense. Each warrants a representation as a percent-
age of revenue. Less frequently we will need to measure sundry, other, and losses/
gains from purchases/dispositions; these only find a place in our margin analysis
bloc if they occur frequently and are meaningful.
      We then present both GAAP pretax income as a percentage of revenue and
PF pretax income as a percentage of revenue. Following are the GAAP tax rate
and PR tax rate—but not as a percentage of revenue. These are represented as
percentages of GAAP and PF pretax income, respectively. We conclude our mar-
gin bloc with GAAP net income as a percentage of revenue (net margin) and PF
net income as a percentage of revenue.
      Once we’ve created the template for these representations for a single quar-
ter, we copy and paste them for all periods: the quarters, and also the half year,
nine months, and full year. In Figure 2.1, the margins for JDSU (formerly JDS
Uniphase) are shown for the fiscal year 2007.


Modeling Forward Periods
To review, we now have at least one and preferably two complete historical years
represented. These years are laid out to accommodate a rolling performance mea-
surement that includes half-year, nine-month, and full-year compilations. A col-
umn enabling year-over-year percentage comparisons now adjoins each quarter
(though not the interim cumulative periods) as well as the full year; in each case,
Figure 2.1

In fiscal 2007, JDSU—which had long struggled to reach adjusted profitability, much less GAAP profitability—had a 4.6% pro forma net margin for the year. On a
GAAP basis, however, net margins were negative.

JDS Uniphase
MARGINS                          1Q07                   2Q07                   1H07        3Q07                  9mos07       4Q07                   2007
COGS % Rvns                        66.1%                  59.8%                  62.7%       62.7%                  62.7%       65.1%                  63.3%
Gross Margin                       30.8%                  37.4%                  34.4%       34.6%                  34.4%       31.9%                  33.8%
Non-GAAP Gross Margin              34.6%                  42.1%                  38.6%       38.8%                  38.7%       35.9%                  38.0%
R&D % Rvns                         12.6%                  11.7%                  12.1%       12.0%                  12.1%       12.0%                  12.1%
SG&A % Rvns                        26.1%                  25.8%                  25.9%       26.5%                  26.1%       27.2%                  26.4%
Amrtzn % Rvns                       2.0%                   1.9%                    1.9%        1.8%                   1.9%        2.0%                   1.9%
Total Operating Expenses           42.3%                  41.7%                  42.0%       41.3%                  41.7%       44.1%                  42.3%
Operating Income                  -11.5%                  -4.2%                   -7.6%      -6.7%                   -7.3%     -12.1%                  -8.5%
Pro Forma Operating Income         -2.9%                   4.1%                    0.8%       -1.7%                 -0.0%       -3.2%                  -0.8%
Interst % Rvnus                    -5.6%                  -4.2%                  -4.8%       -4.0%                  -4.5%        -6.1%                 -4.9%
Pretax Margin                      -5.8%                   7.3%                    1.2%      -2.8%                  -0.2%       -6.4%                   -1.7%
PF Pretax Margin                    2.7%                   8.2%                    5.7%        2.3%                   4.5%        3.0%                   4.1%
Taxes as % Pretax Income            5.9%                  13.8%                  31.0%      -39.2%                 -36.7%       20.4%                  -8.2%
Net Margin                         -5.5%                   6.3%                    0.8%      -3.9%                  -0.8%        -5.1%                  -1.9%
Pro Forma Net Margin                2.1%                   8.2%                    5.4%        3.4%                   4.7%        4.3%                   4.6%
42   •   Income Statement Presentation



the annual percentage-change comparison column is situated immediately to the
left. The income statement presentation has been amended as described previ-
ously to enable quick visual surveys of important figures (gross margin, operat-
ing cost, etc.) in those cases where the company’s own presentation lacked those
tallies. Finally, the presentation has been expanded to accommodate both GAAP
and pro forma results via the addition of several PF-tagged lines. This has been
done to accommodate the (nearly inevitable) major one-time, nonoperating
events that need to be recorded within GAAP results but that would distort our
go-forward assessment of continuing operations.
       For ease of explanation, we’ll create a “model” modeled income statement
with specified cells for each input. For the first time, we will be discussing for-
mulas that reside in certain cells. Our apologies to the Excel-proficient, but we
must begin with the assumption that our modelers are Excel neophytes. As
quickly as we can, we’ll phase down the level of formula detail.
       In transitioning the template of a historical income statement representa-
tion to a living multiquarter model, we will again face numerous chicken-and-
egg moments where we have yet to create inputs that our model needs. In these
instances, we will recommend placeholders in the form of static numbers or
simple percentage changes (reflecting their usual percentage relationship to his-
torical data). With time, we’ll kick out the placeholders and incorporate live
inputs drawn from other parts of the workbook into our income statement
presentation.


Turning the Historical Template into Modeled Quarters
As a first step, copy and paste a full historical model for a single year complete
with static data immediately to the left of the latest historical year. If the model
has been constructed as described previously, all the columns, including the
interim periods and comparison columns, will total 12. In the cells atop the col-
umns of the newly copied model, assign the quarters and interim periods as fol-
lows: 1Q0 n, Y/Y %, 2Q0 n, Y/Y %, 1H0 n, 3Q0 n, Y/Y %, 9mos0 n, 4Q0 n, Y/Y %,
200 n, Y/Y %. I recommend putting an “E” (for estimated) next to all periods.
      We’ll work one quarter at a time. And we’ll assume our historical replicated
years and quarters have pushed us to the right on the spreadsheet as far as column
CA. As a first step, our period designation (i.e., 1Q0 n) resides in cell CA3, total
revenue resides in cell CA4 (or CA6, if the presentation includes product and
service revenue break-outs in CA4 and CA5, respectively), cost of goods is CA5,
and so on.
                            Phase 1: Income Statement and Margin Model, Part 2   •   43



       If you’ve copied the preceding historical year, revenue in quarter 1, 1Q0n,
matches revenue in 1Q0n-1. Later, we’ll refine our forecast based on a granular
assessment; for now, let’s leave it at that. Let’s first model COGS as a percentage
of revenue equal to the preceding full-year average of COGS as a percentage of
revenue. The information on which we base this first formula will be contained
in the historical margin analysis section immediately below the most recent his-
torical income statement presentation. For COGS for period 1, and assuming
total revenue is in cell CA4, our formula is CA4*0.57, where 57% is the annual
average COGS for the prior historical year.
       Isn’t it true that seasonality and other factors influence COGS, resulting in
sometimes predictable variations in gross margin for each quarter? Very true. On
that basis, we could model that first quarter’s gross margin based on the prior year’s
first quarter. But remember, we are in the early stages of efficiently building a rep-
licable model for each quarter; that essential level of refinement comes later.
       When you begin to model seasonal effects in individual COGS, remember
that unlike operating costs, which have a high variable element, COGS contain
many fixed costs. Intuitively, seasonally strong revenue periods will better lever-
age these fixed costs, typically resulting in improved gross margin. But we will
always posit that the best initial guide to seasonal adjustment is to reflect the
experience of prior periods.
       Our hypothetical model includes the three main operating cost categories:
R&D, S&M, and G&A. We will again model these as percentages of revenue
based on the preceding full-year averages. As with COGS, only when the model
is up and running will we amend to reflect seasonal factors.
       Assuming revenue is in CA4 and gross profit in CA6, our three main oper-
ating cost categories are in cells CA7, CA8, and CA9. We will again use the annual
average percentage of revenue shown in the prior year’s margin analysis as our
starting point. If R&D averaged 7.5% of revenue for the prior year, in CA7 we
would use the formula, CA4*.075.
       Even at this simple level of modeling, potential adjustments crowd the
imagination and seek to crowd the page. We would halt the cattle drive dead in
its tracks if we detoured for a discussion of every possible influence on these line
items. But a few are worth mentioning, so at the risk of contorting meaning, we’ll
seek to make these digressions as compact as possible.
       One such topic is incremental margin. After wringing efficiency gains,
companies find that they can generate excess margin on the next, or incremental,
dollar of revenue. As business improves, however, companies need to ramp
spending to match demand, and incremental margin compresses. We’re already
44   •   Income Statement Presentation



seeing that simple, across-the-board percentage of sales inputs won’t do in the
final model. Since we’re already sidetracked, we’ll also take a moment to note that
R&D and G&A are generally less susceptible to seasonal effects than gross mar-
gin, which, with its high fixed-cost component, more meaningfully reflects
swings in the top line. But S&M can have a strong seasonal factor, typically in the
fiscal year’s final quarter. As a rule, if full-year revenues are up nicely, then the
sales staff will receive excess compensation that can increase S&M as a percentage
of revenue in the quarter.
      We’ll be able to accommodate this level of detail more easily in a fully com-
pleted model. In Figure 2.2, for 2009 we highlight Juniper’s actual and modeled
main operating cost items in dollars at the top and as percentages of revenues in
shaded cells at the bottom. Even after a tough first half of 2009, in which sales
were down in the midteens, we assumed that tough comparisons would persist
in 2H09. As of midsummer 2009, we were modeling rising SG&A in 3Q09 to
support new initiatives (mainly Juniper-NSN Carrier Ethernet) and to reflect
sequential top-line gains that would drive compensation. We modeled R&D to
rise as new initiatives took hold, while anticipating that both R&D and S&M
would moderate as such activity slowed in the back half of the year. And we mod-
eled G&A to come down as the company sought cost efficiencies within the cor-
porate office.
      If, and only if, the company’s income statement presentation includes amor-
tization of intangibles, we want to represent this as well—but in relation to a
balance sheet account, not as a percentage of revenue. In our first-phase model,
we use the placeholder of historical four-quarter average amortization. We will
want to tie this in to our balance sheet presentation at a later stage.
      We’ve created lines for the most common nonrecurring or noncash con-
tributors to operating income: acquisition-related, restructuring, impairment,
and the catch-all “other.” Looking forward, it’s a brand-new day, and our mod-
eled company can do no wrong. So we’ll leave these blank for now; most likely
the company will fill them in as it missteps through its operating cycle. Let’s
assign these cells hypothetical values of CA10 through CA13.


Modeling Operating Costs and Operating Income
Our next line represents GAAP operating costs. In cell CA14 on this line, we put
the formula Sum(CA7:CA14). The following line represents PF operating costs.
In cell CA15, we put the formula Sum(CA7:CA9). But we’re not done. If we have
learned that the convention for this company’s pro forma results is to exclude FAS
123R stock option compensation, we proceed as follows.
Figure 2.2

At the time of our writing, we modeled that Juniper would need to ramp some operating costs (principally R&D and sales and marketing) as the Juniper-NSN Carrier
Ethernet joint venture took flight, while it simultaneously sought to contain general and administrative operating costs.

Juniper Networks
Income Statement               1Q09       Y/Y %      2Q09        Y/Y %    1H09        3Q09E      Y/Y % 9mos09         4Q09E       Y/Y %   2009E        Y/Y %
 Product Revenue               587,863       -13%    606,959        -16% 1,194,822     620,762      -19% 1,815,584     634,890       -15% 2,450,474       -16%
 Service Revenue                176,320       19%     179,404        16%   355,724     184,376        2%   540,100     193,595         9%   733,695        11%
 Revenue                        764,183       -7%    786,363        -11% 1,550,546     805,138      -15% 2,355,684     828,485       -10% 3,184,170       -11%
 Prdct COGS                     193,061               207,576              400,637     209,441              610,078    200,328              810,405
 Srvc COGS                       75,451                78,385              153,836      59,073             212,909      56,503              269,412
 Cost of Revenue                268,512        1%     285,961        -1%   554,473     268,514      -13%   822,987     256,830       -16% 1,079,817        -7%
 Gross Profit                    495,671      -11%    500,402        -15%   996,073     536,625      -16% 1,532,698     571,655        -8% 2,104,352       -13%
  Research & Development        185,400        9%     183,894        -1% 369,294       189,208       -2% 558,502        178,124       -1%   736,626         1%
  Sales & Marketing             181,243       -3%     170,575       -10%    351,818    201,285        0%   553,103     190,552        -8%   743,654        -5%
  General & Administrative       39,211       17%       39,175       10%    78,386      36,231       -4%    114,617      33,139      -13%   147,757         2%

 MARGINS
 Product GM                       67.2%               65.8%                  66.5%     66.3%                 66.4%      68.4%                 66.9%
 Service GM                       57.2%               56.3%                  56.8%     68.0%                 60.6%      70.8%                 63.3%
 Gross Margin                    64.9%                63.6%                  64.2%     66.7%                 65.1%      69.0%                 66.1%
 COGS % Rvnus                     35.1%               36.4%                  35.8%     33.4%                 34.9%       31.0%                33.9%
  R&D % Rvnus                    24.3%                23.4%                  23.8%     23.5%                 23.7%       21.5%                23.1%
  S&M % Rvnus                    23.7%                 21.7%                 22.7%     25.0%                 23.5%      23.0%                 23.4%
  G&A % Rvnus                      5.1%                 5.0%                  5.1%      4.5%                  4.9%        4.0%                 4.6%
46   •   Income Statement Presentation



       Down below the bottom of our income statement, we use the information
garnered from historical stock option costs to model average quarterly FAS 123R
costs for the period. The more granular, the better, I always say (up to a point).
If historical information enables identifying and modeling FAS 123R costs by line
item (e.g., COGS and R&D), use the three or four cells needed to do so, and sum
to a single cell; for description purposes, we’ll designate the four line-item cells
as CA36–CA39 and the summary cell CA40. If the information exists as a single
number, model a composite number within the single cell CA40.
       Returning to our PF operating costs line, we amend our formula to read
   Sum(CA7:CA9) CA40, meaning we’re subtracting stock option costs from pro
forma operating costs. Alternatively, we can make an adjustment to percentage
of operating costs. Our analysis of historical precedent for our hypothetical com-
pany may have shown us that, once FAS 123R costs are deducted from the main
operating cost items, these items fairly consistently run at a percentage of main
operating costs (though we may later need to refine this on a per-quarter basis).
       For many technology firms, this is often about 90% of GAAP main operat-
ing costs. For an engineering-intensive firm such as Broadcom, FAS 123R–adjusted
main operating costs can run at a surprising 75% or even 70% of the GAAP
total. For a company whose FAS 123R–adjusted main operating costs tend to run
at 91% of GAAP main operating costs, our formula for PF operating costs
becomes (Sum(CA7:CA9))*.91. As a reminder, we’ll make seasonal adjust-
ments to reflect quarterly variation only after we’ve replicated this quarterly
column across a full four-quarter year. Figure 2.3 shows a snippet of Cisco’s
2009 actual and 1Q10 forecast income statement, with a focus on pro forma
versus GAAP operating costs.
       The next line shows GAAP operating income. For this line, in cell CA16, we
use the formula CA6 CA14, which expresses GAAP gross profit minus GAAP
operating costs.
       For the PF operating income line, we’ll begin with a similar formula,
=CA6 CA15 (GAAP gross profit minus non-GAAP operating costs), modeled in
cell CA17. We did not model PF gross margins, so we can’t use it in our formula.
To square up PF operating income, we need to “muscle up” GAAP gross margin so
it becomes PF gross margin; we do so by adding in FAS 123R costs to GAAP COGS.
I’ve pointed out that this is not a large number, but it is meaningful; for technology
companies, it tends to run a little under 0.5% of revenue. Assuming we modeled
FAS 123R costs to COGS in cell CA36, our amended formula for PF operating
income reads =(CA6 CA36) CA15. Alternately, using the percentage method, we
can use =(CA6*1.005) CA16, assuming COGS-based FAS 123R costs are typically
half a percent of sales, to model PF operating income.
Figure 2.3

This truncated representation of Cisco’s income statement shows actual fiscal 2009 and modeled 1Q10. Amortization (modeled at the prior four-quarter average)
and stock option compensation are the main reasons PF operating costs are much lower than GAAP operating costs.

Cisco Systems
Income Statement              1Q09 Y/Y %        2Q09 Y/Y %         1H09      3Q09 Y/Y % 9mos09         4Q09 Y/Y %         2009 Y/Y % 1Q10E Y/Y %
 Products                       8,635   7.7%     7,347 -10.9%       15,982    6,420 -21.7% 22,402       6,729 -22.1%       29,131 -12.0%  6,842 -20.8%
 Services                       1,696  10.2%      1,742    9.8%      3,438     1,742    9.4%  5,180     1,806     4.8%      6,986    8.5% 1,836   8.3%
   Total Sales                 10,331   8.1%     9,089    -7.5%     19,420     8,162 -16.6% 27,582      8,535 -17.6%        36,117 -8.7%  8,678 -16.0%
 Products COGS                  2,981   5.6%     2,737   -5.0%       5,718    2,327 -18.8%    8,045     2,436 -20.4%       10,481  -9.9%  2,446 -17.9%
 Services COGS                    669  19.9%        629    3.3%      1,298       606  -2.4%   1,904        638    0.2%      2,542    4.8%   652  -2.6%
   Total COGS                   3,650   8.0%     3,366   -3.6%       7,016    2,933 -15.9%    9,949      3,074 -16.9%      13,023   -7.3% 3,098 -15.1%
  Total Gross Profit             6,681   8.2%     5,723   -9.7%      12,404    5,229 -17.1% 17,633       5,461 -18.1%      23,094   -9.4%  5,580 -16.5%
R&D                             1,406  18.0%      1,279    5.2%      2,685     1,243 -13.6%   3,928     1,280   -2.0%       5,208    1.1% 1,237 -12.1%
S&M                             2,283  14.0%      2,155    3.4%      4,438    1,956    -8.1%  6,394     2,009    -7.2%      8,403    0.3% 1,974 -13.5%
G&A                               395 -19.4%        380 -26.9%         775       302 -37.0%   1,077        488  -5.8%       1,565 -22.0%    456  15.3%
Pyrl Tx Stk Optn Excrs              -   0.0%          -    0.0%          -         -    0.0%      -          -    0.0%           -   0.0%     -   0.0%
Amrtzn Dfrd Stk-Bsd Cmpnsn          -                 -                  -         -              -          -                   -            -
Amrtzn Intngbls                   112  -4.3%        136  17.2%         248       121    3.4%    369        164  10.1%         533    6.8%   133  18.8%
In-Process R&D                      3   0.0%          -    0.0%          3         -    0.0%      3         60    0.0%          63   0.0%     -
Acquisition Costs                   -                 -                  -         -              -          -                   -            -
    Oprtng Expenses             4,199  10.4%     3,950     0.4%      8,149    3,622 -13.0%   11,771      4,001  -3.3%      15,772         3,800  -9.5%
    PF Oprtng Expenses          3,824  11.6%      3,515   -1.6%      7,340     3,146 -15.2% 10,662      3,379   -8.2%       14,181        3,271 -14.5%
    Oprtng Incm                 2,482             1,773              4,255     1,607          5,862     1,460               7,322         1,781
    PF Oprtng Incm              2,898   3.5%     2,246 -19.9%        5,145     2,125 -20.5%  7,094       2,129 -29.6%       9,083 -19.3%  2,356 -18.7%

                                                                                                                                                  (continued)
Figure 2.3 (continued)


Cisco Systems
Income Statement              1Q09     Y/Y %    2Q09     Y/Y %    1H09     3Q09     Y/Y % 9mos09     4Q09     Y/Y %    2009      Y/Y %   1Q10E    Y/Y %
  FAS 123R:
COGS-Prdcts                       15     0.5%       14     0.5%       29       14     0.6%      43       13     0.5%        56               13     0.5%
COGS-Srvcs                        27     4.0%       25     4.0%       52       28     4.6%      80       34     5.3%       114               34     5.2%
                                  42     1.1%       39     1.2%       81       42     1.4%     123       47     1.5%       170               47     1.5%
R&D                               95     6.8%       86     6.8%      181       89     7.2%     270       85     6.6%      355                84     6.8%
S&M                              131     5.8%      124     5.8%      255      118     6.0%     373      120     6.0%      493               119     6.0%
G&A                               34     8.5%       32     8.5%       66       28     9.3%      94       53    10.9%       147               52    11.5%
  Subtotal                       301               281               583      277             860       305              1,165              302
Cmpsnation Expns Acqstns         102               101               203      120             323       100               423               100
Higher Amortization               40                40                80       40              120       40                160               40
Tax Effect                      (84)              (79)             (163)    (156)            (319)     (85)             (405)              (69)
   Total Aftr-Tx GAAP to PF      359               344               703      281             984       360             1,343               373
   P&L Aftr-Tx GAAP to PF        292               381                        394                       703                                 443
                           Phase 1: Income Statement and Margin Model, Part 2   •   49



Interest Cost and Interest Income: Preliminary
We now need to sum for GAAP and pro forma pretax income. But first, having
modeled operating costs and income, we need to model those nonoperating items
that impact pretax income.
       Interest income and interest costs are considered nonoperating for a rea-
son: they are not a function of operations, but of financial strategy and deci-
sions. For this reason alone, interest income and interest cost should not be
modeled as a percentage of revenue. Intuitively, we know that for companies
with a strong seasonal component to their quarterly reporting (e.g., retailers
and consumer electronics manufacturers), percentage-of-revenue estimates
for relatively stable balance sheet–based items would oscillate unrealistically
from quarter to quarter and lead to inaccurate modeling of these important
line items.
       To model interest costs or interest income earned, we need to import infor-
mation on cash on hand and debts owed from the balance sheet. For cash, these
most typically include cash and equivalents and short-term investments from
current assets, and long-term investments from long-term assets. For debt, this
includes short-term debt and/or long-term debt due within one year from current
liabilities, and long-term debt from long-term liabilities.
       In our chicken-or-egg quest to get to value, however, we haven’t modeled a
balance sheet yet. Rather than link to a cell on our as-yet-nonexistent balance
sheet, we’ll use as a placeholder the most recent historical quarterly numbers
entered as static values.
       In the very early stages, we use placeholders for interest cost and interest
income. We express interest expense as a positive number, but we express interest
income as a negative number. This may seem a bit counterintuitive, but it is con-
sistent with our presentation. All our line items below revenue so far have been
costs against revenue expressed as positive numbers; interest income is the first
offset to those costs. Note that U.S.-based companies operating under U.S. GAAP
present income statement costs as positive numbers, while foreign companies
operating under IFRS present cost line items as negative numbers.
       The placeholder for both interest income and interest cost is the most recent
quarterly number, typically 4Q of the preceding historical year. Note that we are
not using the average for the full preceding year. The income statement shows a
company’s progress across a given period, while the balance sheet is our most
timely measure of a company’s financial status right now. Over the preceding
year, the company may have used up cash, reducing its interest income line, or
paid off debt, benefiting its interest income line.
50   •   Income Statement Presentation



      We put our roughly modeled interest cost in cell CA18 and interest income
in cell CA19. We have also reserved two lines: for (1) sundry nonoperating other
and (2) gains (or loss) on investments. In our modeled template quarter, these
occupy CA20 and CA21, both of which we leave blank.


Interest Cost and Interest Income: Adjustments
Later in our process, we’ll model the balance sheet and cash flow statement. Once
the balance sheet for forward periods is modeled on a worksheet called Ratios
and Valuations (R&V), we need to link to this sheet to complete our interest
income and interest cost models.
       All the balance sheet line items related to cash, cash equivalents, short-term
investments, and (often) long-term investments will contribute to cash and will
contribute to interest income. On the top of our balance sheet annual compila-
tion on the R&V worksheet, we sum all these accounts. All the balance sheet line
items related to short-term debt, commercial paper, long-term debt due within
one year, and long-term debt will impact interest costs. We sum these atop the
liabilities section in our annual balance sheet presentation. (We italicize both
summed cash and summed debt so they stand out from the balance sheet
itself.)
       At the bottom of the income statement presentation, we’ve left room for
items not purely part of the actual income statement (e.g., dividends and stock
option compensation) but that either impact the pro forma calculation or will be
useful later on. Immediately below dividends and above FAS 123R (stock option
compensation), insert a line for cash and investments and a second line for total
debt. For our example, let’s assume cash and investments is in row 50 and total
debt is in row 51.
       In the column for the first quarter of any modeled year, link the annual cash
tallied on the balance sheet to the cash and investments row in cell CA50; and
link the summed debt from the balance sheet to the total debt row in cell CA51.
For the remaining three quarters of that year, link these tallies with a simple
equal sign in rows 50 and 51. There is no need to reference the annual cash and
debt in the interim-period columns.
       The company’s own representation of interest income, interest cost, and/or
net interest cost (income) will drive our modeling procedure. The company that
provides both interest income and interest cost in its quarterly presentation is
easiest to work with. For the first quarter, in the interest cost cell CA18, put in
the formula, CA50*.25*.07; in this formula, CA50 is the debt to be serviced,
the .25 represents one-quarter’s interest obligation, and the .07 is a hypothetical
                           Phase 1: Income Statement and Margin Model, Part 2   •   51



7% interest rate that about hits the corporate average coupon. This formula will
produce a value that will have some relationship to the nearest historical quarter’s
interest cost.
      Now the calibration begins. In advance of this exercise, using data from the
10-K or Q you can weight the company’s debt adjusted by interest cost and pro-
duce your own estimate. But we find this number is often not a reliable predictor
of income statement interest cost. The quicker procedure is try on different inter-
est rates until you get the rate that best matches the nearest historical quarter.
Remember, quarterly debt payments have (almost) no seasonal component. At
each quarterly report, changes in balance sheet debt will change the interest cost
estimate. But you must track the nature of those changes; for example, if a 5%
coupon instrument is rolled over into a 9% bond, your modeled quarterly interest
rate must adjust upward.
      The interest income process and calculation is similar, with a few signifi-
cant differences. First, yield on cash is rarely more than 3% and often a great deal
lower. During the worst days of late 2008 and early 2009, CFOs were telling us
they were moving their cash “closer to the printer” (i.e., buying only Treasury
debt, which at the time had fractional yields). Second, companies tend to lard in
other and sundry items into interest income; in our experience it is rarely as clean
a number as interest expense. In modeled quarter 1, the formula for interest
income in cell A19 would be =CA51*.25*.0275, where our return on cash
assumption is about 2.75% annually. When you calibrate this cell, you almost
always model the rate down.
      For companies providing net interest income (cost) only, and assuming our
initial assumptions were correct and confirmed by the calibration, the formula
in CA18 would be (CA50*.25*.07) ( CA51*.25*.0275). Always, we want to
model these line items with links to the financial structure information. But we
also want to accommodate information from the company, such as news that the
board will engage in early debt retirement.


Pretax Income
In cell CA22, our formula for GAAP pretax income is =CA16 (SUM: CA18:CA21).
We will not be adjusting any of the GAAP line items between GAAP operating
income and GAAP pretax income.
      We generally do not need any overt adjustments to net interest expense
within our standard formula for PF pretax income. The formula for PF pretax
income in cell CA23 is =CA17 (SUM: CA18:CA21). Companies will sometimes
52   •   Income Statement Presentation



provide GAAP versus adjusted net interest expense information. Qualcomm is
an example of a company that provides this level of information; the discrepancy
mainly results from its pro forma treatment of an entire business unit (i.e., Qual-
comm Strategic Initiatives). Figure 2.4 shows a snippet of Qualcomm’s 2006 P&L,
with distinctive values for GAAP and pro forma interest income. Oddly, the two
columns roughly square up by year end.
      In their post-release presentations, some companies may also call out items
in “other” or in gains/loss on investments that they feel should be excluded from
pro forma consideration. Although such callouts are reasonably common, they
are irregular enough that we will not allot a line item for this purpose.


GAAP and PF Taxes
In cell CA24 we model GAAP income tax expense. Income tax expense is a per-
centage-of calculation, but not percentage of revenues; companies are taxed on
their pretax earnings, so we’ll use percentage of pretax income.
      I’ve always found GAAP taxes to be an elusive animal. Particularly in the presence
of nonoperating events, GAAP taxes move in mysterious ways. Impairments frequently
trigger charges for tax valuation allowances; even amid big GAAP losses, a company
may report paying GAAP taxes. As discussed earlier, this can reflect global companies
making good money in some nations and losing money in others.
      Corning, for example, is structured to be unprofitable in the United States.
Its nonpareil R&D business, largely U.S.-based, is unmatched in turning “blue-
sky” research into hard dollars. But the products issuing from Corning’s research,
such as precision glass, are mainly sold in Asia and other markets. It is not easy
getting Corning’s GAAP tax mix just right, even with guidance from the com-
pany. Nonetheless, we’ll model as best we can using historical precedent and/or
direct information from the company. For our model, and assuming historical
information (or company guidance) has pointed to a GAAP tax rate averaging
34%, in cell CA24 we use the formula =CA22*.34.
      Pro forma taxes, by contrast, are often set by the company and end up being
a fairly consistent percentage of PF pretax income, come rain and come shine.
Given that adjusted profits increasingly drive investment decision making, in
topsy-turvy thinking that (scarily) might make sense companies reason that pro
forma results, and thus pro forma taxes, should be reliable—even though they’re
a shared hallucination. Ours is not to reason why . . . again using historical prec-
edent and/or direct information from the company, and assuming an average
28% rate, we model PF income tax expense in cell CA25 with the formula
=CA23*.28.
Figure 2.4

Qualcomm is one of a handful of companies that reports (or at least shares with analysts) both GAAP and adjusted interest income. However you slice it, this
company earns a lot on its cash horde.

 Qualcomm
 Incm Stmnt (Sept)          1Q06        Y/Y %     2Q06       Y/Y %     1H06       3Q06        Y/Y %  9mos06        4Q06        Y/Y %     2006        Y/Y %
 Eqpmnt, Servcs              1,150.0               1,122.0      32%     2,272.0    1,240.0       41%   3,677.7      1,264.0               5,015.7       31%
 Licensing, Royalty Fees       591.0                 712.0      38%     1,303.0       711.0      49%   1,849.0        735.0               2,510.0       37%
 Total Revenue               1,741.0      25%      1,834.0      34%     3,575.0    1,951.0       44%  5,526.7       1,999.0      28%      7,525.7       33%
   Cost of revenue              517.0     20%        521.0      35%     1,038.0      559.0       44%   1,597.0        586.0      33%      2,183.0       33%
   R&D                        340.0       55%        390.0      55%       730.0      395.0       61%   1,125.0         411.0     61%      1,536.0       58%
   SG&A                        239.0      67%        263.0      70%       502.0      293.0       94%     795.0        321.0      91%       1,116.0      81%
   Amortization                     -                    -                    -           -                   -            -                     -
   Prchsd In-Prcss R&D                                                        -                               -                                  -
   Other/Asst Imprmnt                      0%                    0%           -                   0%          -                   0%             -      0%
     Total Oprtng Expns      1,096.0      38%      1,174.0      48%     2,270.0     1,247.0      56%    3,517.0     1,318.0      48%      4,835.0      49%
     PF Oprtng Expnss          948.0               1,001.0              1,949.0    1,084.9            3,033.9       1,153.3                4,187.2
 Op Incm P&L                   645.0       8%       660.0       15%     1,305.0      704.0       26%  2,009.7         681.0       2%      2,690.7       10%
 PF Op Incm                    793.0                833.0               1,626.0      866.1             2,492.1       845.8                3,337.9
   GAAP Interest Incm            91.0      0%        125.0       0%       216.0      120.0        0%     336.0        129.0        0%       465.0        0%
   PF Interest income           111.0     68%        129.0     111%       240.0       110.0      31%     350.0        114.0       -3%       464.0       41%
   Invst Loss/Dstrbtns              -      0%                    0%                               0%          -                    0%            -       0%
   Other                                                                                                      -                                  -
  Pretax Incm P&L             736.0        5%       785.0       24%     1,521.0      824.0       20%  2,345.7         810.0       3%      3,155.7       14%
 PF Pretax Incm P&L           904.0                 962.0               1,866.0      976.1             2,842.1        959.8               3,801.9
54   •   Income Statement Presentation



      In cell CA26, we express GAAP taxes as a percentage of GAAP operating
income; in cell CA27, we express pro forma taxes as a percentage of PF operating
income. I typically italicize these lines to make them jump out, and these are the
only percentages within the quarterly columns. Yes, we replicate these percent-
ages in our margin presentation below. Remember that our model is about visual
reference and quick trend identification, with pertinent items in their logical
places.
      The next step is to tally up GAAP and PF net income before various “below-
the-line” items. In cell CA28 in our hypothetical Q1 model, our header is “Net
Income before MI, EE, DC, & D”; these acronyms are discussed both above, in
our historical section, and below. The formula for cell CA28 (GAAP net income
before MI, EE, DC, & D) is CA22 CA24. The formula for cell CA29 (PF net
income before MI, EE, DC, & D) is CA23 CA25.


GAAP and PF Net Income
We’re almost ready to model net income, but first we need to model the final
items that may impact it. We’ll designate four cells for equity income (CA30),
minority interest (CA31), discontinued operations (CA32), and accounting
changes (CA33).
      So far we’ve expressed every cost against revenue as a positive number, and
every add-back (e.g., interest income) as a negative number. That accords with
most companies’ presentation style. Below pretax income, however, many com-
panies will present costs to income (e.g. minority interest and losses from discon-
tinued operations) as negatives; and they treat add-backs to income (e.g., equity
income and income from discontinued operations) as positives.
      In our style, we will use the same convention throughout. Minority interest
(assuming it represents disgorgement of income and not sharing in a loss) and
losses from discontinued operations are expressed as positive numbers; equity
income (assuming it is positive) and income from discontinued operations are
expressed as negative numbers.
      Although they are straightforward to model, equity income and minority
interest are not easy to model reliably. As our model grows in sophistication, we
can include information on equity and minority partners; that will give us a more
granular basis for predicting these inputs. Remember that equity income is the
portion of income allotted to a 20%-to-50% equity partner. Beginning with the
base assumption that the joint venture is profitable, we would model this input
as a negative number. Minority interest expresses that portion of JV income dis-
                           Phase 1: Income Statement and Margin Model, Part 2   •   55



gorged or not retained by the majority partner. Because it is in effect a “cost,” we
will model it as a positive number.
      Until our model gains the necessary sophistication, we will use the one-year
average value for equity income and for minority interest. We will express minor-
ity interest not as a percentage of revenue, but as the historical average percentage
of pretax income based on the preceding four quarters. Equity income is a bit
trickier, as it is coming from an off-screen partner. In the preliminary stages, and
until we know more about this partner (and the partnership), we will also model
equity income as a four-quarter average.
      The last two items, discontinued operations (CA32) and accounting changes
(CA33), we can leave as blanks for now.
      Cell CA34 in our hypothetical model represents GAAP net income. We’ll
use the formula, CA28 (Sum:CA30:CA33).
      Cell CA35 in our hypothetical model represents PF net income. In general,
and there may be exceptions, we will include minority interest and equity income
in our PF calculation, but not accounting effects or income/losses from discon-
tinued operations. In cell CA36, our formula for PF net income is CA29
(Sum:CA30:CA31).


Per-Share Earnings
The next two lines are devoted to basic and diluted shares; we’ll be modeling
them in CA37 and CA38. Up above we discussed, in capsule form, the distinction
between basic and diluted share count. For both, we will model off the nearest
historical period, not the year-earlier period or the full-year average.
      One of the reasons that we include pro forma calculations as a basic part of
our model (the other being the chance/likelihood of catastrophic impairments)
is because of share-issuance policy. All public companies use shares as compensa-
tion to one degree or another. Information industry companies tend to have high
levels of share compensation; manufacturing companies tend to have lesser share
compensation. To put it more graphically, companies whose employees shower
before work issue lots of stock as compensation; companies whose employees
shower after work issue less stock as compensation.
      Companies will also publicize their intention to buy shares (although they
are typically less forthright when they disengage from this practice). Through
perfectly permissible accounting devilry, a company may use cash to repurchase
shares; this is treated as a financial, not an operating event. These costs do not
run through the income statement. But they reduce the share base and the
56   •   Income Statement Presentation



denominator for EPS calculations, thus giving the impression of operating
improvement.
      This perfectly legal accounting sleight of hand, as far as I can tell, is the
main reason companies pour cash down the sinkhole of share repurchase.
Granted, companies with high dividend payouts may want to reduce their income
obligation by shaving the share base. Many companies will choose to repurchase
sufficient stock to offset, to one degree or another, stock issued as
compensation.
      The worst offenders are often dividend-free and are buying Treasury stock
well in excess of their stock-compensation issuance. Repurchased shares are
recorded in the “Treasury” account in stockholders’ equity, thus improving
return on equity. Because the corollary effect is to boost the debt/capitalization
(debt/cap) ratio (see Chapter 9), this practice is usually (though not always) lim-
ited to low-debt companies. In the “justice served” department, at the depth of
the market downturn during spring 2009 dozens of companies were identified as
having market capitalizations that were less than the amount of money they had
spent to repurchase stock in 2008.
      A cursory examination of the preceding eight quarters’ historical share base
will indicate the direction of the outstanding share count. Unless recent policy
has shown us otherwise, we will model some “creep” in the share base, perhaps
a quarter-percentage point per sequential quarter. If the company has historically
been a repurchaser, we’ll assume some modest decline in quarterly share base.
      Assuming the share base has historically crept higher and that column BW
contains the most recent historical period, our formula for basic shares outstand-
ing in cell CA37 is BW37*1.0025. This formula indicates that the share base
will grow one-quarter of a percentage point for the quarter. Drag and drop this
formula down one cell to CA38 for diluted shares outstanding; thus,
  BW38*1.0025.
      With our share bases now established, we can finally model earnings per
share, the single most important number in most valuation calculations (for bet-
ter or worse). In cell CA39, GAAP basic EPS, we use the formula CA35/CA37,
where CA35 is GAAP net income and CA37 is basic shares outstanding.
      For GAAP diluted EPS, as a starting point in cell CA40, we can use a similar
formula, CA35/CA38, where CA35 is GAAP net income and CA38 is diluted
shares outstanding. For most companies, particularly those whose basic and
diluted share bases are similar (within 1% to 2% of one another), this will be
sufficient.
                           Phase 1: Income Statement and Margin Model, Part 2   •   57



      There is, however, an important potential variable in the calculation of
GAAP diluted EPS (and in pro forma diluted EPS). Common stock equivalents
(CSEs) can affect both the numerator and the denominator in our diluted EPS
calculations. In the snuggest possible nutshell, the CSE that most notably affects
diluted net income and share count is convertible debt. When a company is prof-
itable, the number of shares that would be represented assuming the debt were
converted into common must be added to the denominator. If we are assuming
that this is no longer debt but equity, we need to pull out that portion of interest
paid related to the convertible debt turned converted shares and add it back to
income.
      What does this mean to the modeler trying to herd cattle rather than wran-
gle with theory? Begin by surveying the prior periods. In our experience, a wide
gap between basic shares and diluted shares—say, more than 3%—is often a flag
for the presence of convertible debt. (We’re assuming it is early enough in the
process that we haven’t seen much of the balance sheet.) You can and should go
through the 10-K and the 10-Qs, and that’s never a bad idea, for data on the con-
vertible debt.
      In Figure 2.5, we see the variability in Ciena’s share base for the October
2007 year. While basic shares mainly stayed near 85 million in all four quarters,
diluted shares outstanding rose much higher. Accordingly, basic share count
increased 2% on average for the year, while diluted share count rose 11%. Ciena
ended fiscal 2007 with more than $500 million in short-term convertible debt
and $800 million in long-term convertible debt. “Convertible” debt is debt that
at maturity can convert into common shares; the common shares represented
within the convertible can potentially be counted as CSEs. The more profitable
the quarter, the higher the diluted share base as more of the CSEs qualified for
conversion.
      More prosaically, for companies with wide variation between basic and
diluted shares, use the starting point formula, CA35/CA38, (where CA35 is
GAAP net income and CA38 is diluted shares outstanding) for four to eight
historical quarters. See if the formula consistently renders a number different
from reported GAAP diluted EPS; if so, that degree of variation is a guideline for
adjusting for the convertible debt. It is important to perform this test with GAAP
diluted EPS, because it always is reported; pro forma diluted EPS might be a
footnote in the press release.
      If you consistently determine a one-penny difference between what is ren-
dered by the formula and what the company reports, then you can treat one
Figure 2.5

This snippet from Ciena’s fiscal 2007 income statement shows that basic shares outstanding stayed consistent near 85 million across the four quarters, while the
diluted share base ranged from 93 million to 109 million. The company had over $1.3 billion in convertible debt at fiscal year-end 2007. Note that the higher the
net income per quarter, the higher the diluted share base.

 Ciena
 Income Statement             1Q07                  2Q07         Y/Y %      1H07        3Q07       Y/Y %   9mos07      4Q07        Y/Y %     2007       Y/Y %
    Operating expns             70.84         8%      79.09           13%    149.93       81.64        -3%   231.57      82.03         19%    313.59         8%
    PF Operating expns          61.82        -1%      68.04            8%    129.86       72.29        14%   202.15      77.48         29%    279.62        13%
     Operating income             2.79     -118%       2.74         -138%      5.53       16.03      -225%    21.56       27.12      599%      48.68      -255%
     Pro Forma Oprtng Incm       11.81     -201%      13.79       13287%      25.60       25.38       211%    50.98      31.67        158%     82.65       782%
 Purchased R&D
 (Loss) Gain Mrktbl Debt
 Interest/other incm            14.85       60%        16.90         51%       31.74       19.46      39%      51.21      25.28       61%       76.48       52%
 Interest expens                (6.15)       2%        (6.15)         6%     (12.30)      (6.93)      13%    (19.23)      (7.77)      26%     (27.00)       12%
 Lss on Invstmt, Dbt Extng           -                       -                      -       0.59                0.59    (13.01)               (12.42)
    Pre-tax income               11.49     -292%       13.49       -976%       24.98       29.15     -835%     54.13       31.62     135%       85.74     4239%
    PF Pretax Incm              20.51      -539%       30.03         218%      53.74       42.16      112%     95.90       52.18     110%      148.07      196%
 Income Tax                       0.42       41%         0.48         29%       0.90        0.84      163%      1.74        1.21     207%        2.94      113%
 PF Incm Tx                       0.42     -127%         6.01         93%        6.43       0.84      -87%      7.27        1.21     -88%        8.47      -52%
Tax rate                           4%      -173%          4%        -115%         4%         3%      -136%       3%          4%       31%         3%       -95%
PF tax rate                        2%       -94%         20%         -39%       12%          2%       -94%       8%          2%      -94%         6%       -84%
 Net income                      11.07     -276%        13.01      -781%       24.08       28.31     -761%     52.39      30.41      132%       82.80    13816%
 Pro forma net income          20.086      -743%       24.02        280%        47.31      41.32      204%     88.63      50.97      241%      139.60      333%
 Avg Basic Shares               84.95         2%       85.20           2%      85.08       85.56        2%     85.24      86.24        2%       85.49        2%
 Avg diluted shares             93.26        12%       93.74           7%      93.50      101.57        9%     96.19     108.81       17%       99.34       11%
 Reprtd Basic EPS                 0.13     -272%          0.15     -768%         0.28       0.33     -750%      0.61        0.35     128%        0.97    13547%
 Rprtd Dilutd EPS                 0.12     -257%          0.14     -736%         0.26       0.28     -705%      0.54        0.30     113%        0.87    12985%
 Pro forma Basic EPS              0.24     -727%         0.28        272%       0.56        0.48      199%      1.04        0.59     235%        1.63      325%
 Pro forma Dltd EPS               0.22     -672%         0.26        254%        0.51       0.41      178%      0.92        0.48     198%        1.41      289%
                           Phase 1: Income Statement and Margin Model, Part 2   •   59



penny as a fairly reliable add-back. I can hear the purists gnashing their teeth. We
are going to designate a line below our other imported inputs (e.g., cash, debt,
intangibles). We call this line CSE interest add-back. And in cell CA50, we put in
the calculated difference, in this case $0.01, as a static number.
      Moving forward, in cell CA40 our amended formula for GAAP diluted EPS
is (CA35/CA38) CA44, where CA35 is GAAP net income, CA38 is diluted
shares outstanding, and CA44 is our per-share-based average interest add-back.
      If a company has income and/or loss from discontinued operations to
report, it will be reported, both as an absolute number and on a per-share basis.
To prepare for this eventuality (no company avoids it forever), we will create
blank placeholder cells CA41 (basic discontinued ops EPS) and CA42 (diluted
discontinued ops EPS).
      Because we’ve been so careful to distinguish GAAP from pro forma results
all through the income statement, we won’t need any awkward plug-in number
given by the company and taken on faith. For pro forma results we’ll mainly
replicate the two formulas used for GAAP per-share results. In cell CA43, pro
forma basic EPS, we use the formula CA36/CA37, where CA36 is PF net income
and CA37 is basic shares outstanding.
      For pro forma diluted EPS, and assuming that convertible debt or other CSEs
have no impact on per-share results, in cell CA44 we can use a similar formula,
   CA36/CA38, where CA35 is PF net income and CA38 is diluted shares out-
standing. If we have identified and noted the CSE interest add-back and recorded
it in cell CA50, our cell CA44 amended formula for pro forma diluted EPS, is
   (CA36/CA38) CA50, where CA36 is PF net income, CA38 is diluted shares
outstanding, and CA50 is our per-share-based average interest add-back.
      As a reminder, we’ll include the quarterly dividend immediately below PF
diluted EPS, even though this is not part of the formal income statement. Use the
immediately preceding historical quarterly payout as a placeholder, unless you
have information about a change in dividend policy.


Turning One Column into a Full-Year Model
Remember we have a 12-column full-year model, which encompasses the four
quarters and the interim periods (half year, nine months, and full year), along
with five year-over-year percentage-change columns. Consistent with our histori-
cal period presentation, to distinguish the year-over-year comparison columns
from modeled quarters and rolling summary periods in modeled periods, itali-
cize the Y/Y comparison columns. These five columns get an identical header:
Y/Y %.
60   •   Income Statement Presentation



      For all but the Y/Y comparison columns, copy and paste an exact replica of
our freshly modeled quarter onto the remaining columns, including the rolling
summary periods. We’ll now proceed to make the necessary adjustments, first in
the modeled quarters and then in the rolling summary periods.
      The copying and pasting of Q1 onto subsequent quarters is fairly straight-
forward. However, when you use 1Q as a template to replicate Q2, Q3, and Q4,
remember that the nonsymmetrical staging of periods means you must be mind-
ful of values drawn from preceding quarters. This has bearing, for example, in
the share count, which needs to reference the nearest period. For example, we
modeled the Q1 share base off the preceding 4Q, which is four columns distant.
But 2Q will model its share count off 1Q, which is two columns distant. If you
simply copy and paste values here, the source column will be not 1Q, but a full
year. Be mindful of needed amendments to link to the proper period.
      For the rolling summary periods, the adjustments are not complicated, but
they take some care. And generally, we will make distinctions between rolling
summary compilation of GAAP line items compared to pro forma line items.


Compilation Columns (Half Year, Nine Months, and Full Year)
In our half-year rolling summary, we will generally sum reported line items from
1Q and 2Q. For GAAP formula cells, we will retain single-column summary
formulas down the column for line items such as GAAP operating costs and GAAP
pretax income. For pro forma formula cells, we will most typically sum the results
of individual quarterly cells for Q1 and Q2 across the page. Let’s first talk most
broadly about the half year, then extrapolate those practices to the nine months
and half year.
      Walking down the income statement, on the premise that revenue is in CA3
for 1Q and in CC3 for 2Q, in cell CE3 use the formula CA3 CC3. When you
get to formula cells, for instance gross margin, retain this across-the-page sum-
mary formula. After summing the operating cost line items across the page,
retain the down-the-column formulas for GAAP operating costs and GAAP
operating income. We typically do not retain the formula for PF operating cost
and operating income; instead we sum the two quarterly values across the page.
Companies are not always fully transparent about what they include in their pro
forma calculations; and as the interim periods progress from half year to nine
months to a full year, small discrepancies are amplified.
      Interest income, interest expense, gains and losses on investments, and
other are summed. Similar to our treatment of operating cost and income, while
                          Phase 1: Income Statement and Margin Model, Part 2   •   61



we retain the down-the-column formula for GAAP pretax income in the half-
year column, for PF pretax income we sum the two quarterly values across the
page. Once pro forma is acknowledged by a company, line items such as interest
income can be adjusted in unpredictable ways.
      Like the individual operating cost line items, taxes are summed across the
page, both GAAP and PF. The summed tax rate as a percentage of half-year pre-
tax income remains a formula. I retain the down-the-column formula for GAAP
net income, and I use the across-the-page formula to sum PF net income. For
basic shares outstanding for the half year, in column CE37 we use the formula
  (CA37 CC37)/2, representing the average of quarters 1 and 2. For diluted
shares outstanding, drag and copy this formula into cell CE38.
      In cell CE39 for GAAP basic EPS, retain the formula. For GAAP diluted EPS,
however, remember that convertible debt and other CSEs can alter both the numer-
ator and the denominator in unpredictable ways. Therefore, in cell CE40 for GAAP
diluted EPS we sum the two quarterly values. In cells CE41 and CE42, for discon-
tinued operations, we sum the placeholders (zero plus zero equals zero).
      In cell CE43 for pro forma basic EPS, retain the down-the-column formula.
For pro forma diluted EPS, in cell CE44 we sum the two quarterly values across
the page, again because convertible debt and other CSEs can alter both numera-
tor and denominator.
      Conclude by summing the two quarterly dividend payments, which though
not part of the income statement are part of our income statement presentation.
      For the nine-month column, repeat the process as directed for the half-year
column. Again, for multiquarter periods, sum GAAP inputs down the column; and
sum pro forma inputs across the page (e.g., 1H plus 3Q). Note that because of the
nonsymmetrical spacing of the quarterly columns, you cannot copy and paste the
half-year column onto the nine-month column; you will need to adjust all values.
For the basic and diluted share base, you must average three quarterly periods.
      You can, however, cut and paste the nine-month column to create the full-
year column—assuming you have strictly replicated the 12-column full-year
scheme I have delineated. The essential adjustment after cutting and pasting the
nine-month column onto the full-year column is in basic and diluted shares
outstanding, which must now average all four quarters.
      You have now, incompletely, modeled a full-year income statement. You
could cut and paste this model and begin on the following year. But that is pre-
mature, from our perspective, because the model is static, unresponsive to mar-
ket realities, and not at all attuned to the company’s prospects. Not just revenue
but the individual cost line items are unrealistic.
62   •   Income Statement Presentation



      Before we can begin inputting seasonality and other real-world flavors into
the individual quarters, we need to make the top line more reflective of market
realities.
      We’ve now completed the two parts of our first phase: (1) building, or rather
rebuilding, the income statement structure for real-world analysis and (2) model-
ing, at least on a preliminary basis, the inputs that will get us from top line to
bottom line. In our next chapter, we’ll discuss means by which we can make that
top line more informed and more nuanced, in pursuit of a more accurate bottom
line.
                                                     Chapter 3
           PHASE 2:
           SEGMENT MODELING
           OF REVENUES




Segment Modeling: A Discussion
Let’s take a real-world situation, and a common one at that. A company announces
that its mobile-device unit sales in Europe were “disappointing” in the just-con-
cluded quarter, particularly for high-end devices. The press release is scant. Calls
to the company are met with the official brick wall known as Reg FD [Regulation
Fair Disclosure is designed to level the investing playing field by compelling IR
departments to share the same level of information with all investors. Given the
complications attendant on universal disclosure, the practical effect of Reg FD
has been to prompt companies to reveal less and less information. (see Reg FD).]
What to do?
       Standard-issue top-line modeling consists of using the historical annual
percentage growth pertinent to that period; let’s say it is 7%. We just found out
about the lousy mobile phone sales, and we need to do something. Do we shave
that forecast growth to 6.75%? How about 6.875%? What do we do with subse-
quent periods? The inadequacy of the answers suggests that any such change is
guesswork, with a bad input likely to pay compound interest over time.
       It is possible to build a multistage revenue model that can reflect a better, if
not perfect, representation of that announced effect not just on current period
revenues but on go-forward modeling. Equally important, our overarching goal
                                                                                    63
64   •   Income Statement Presentation



is to create a model whereby each such change in revenue assumptions ripples up
the income statement, out across the valuation worksheets, and into our final
dollar-based calculation of asset fair value. First, though, we need to refine our
revenue growth assumptions.
       The goal of segment-based modeling is to model revenues as precisely as
possible. We don’t ever want to assume that revenues in aggregate will grow by
some predictable percentage relative to prior-period aggregate numbers, simply
because that has been the experience in the past.
       To model revenue in the real world, we need to work with what companies
give us, what competitor companies are saying about the environment, and all
other available sources. We always begin with what companies are saying offi-
cially about their business.
       For U.S.-based publicly traded companies, the Securities and Exchange
Commission (SEC) has mandated that revenue contributions be identified when
they reach certain thresholds as a percentage of revenue. This division may be
done on a product-line basis or on a geographic basis.
       In the past, public companies often fought to keep their segment data pri-
vate for competitive purposes. Companies gradually came to recognize that their
competitors had sufficient research capabilities and were able to collect this infor-
mation regardless of efforts to safeguard such information. The investing public,
meanwhile, was being left uninformed by such secrecy. Companies have further
realized that the some level of disclosure prompts increased brokerage coverage,
which in turn raises their profile before investors. We would say that, compared
to prior decades, companies are no longer hoarding information that their com-
petitors have had anyway.
       In providing more detailed disclosure, companies can highlight what’s
going right (e.g., “We’re winning in printers!”) while downplaying the negative
(e.g., not mentioning poor sales of mobile phones). Interestingly, more detailed
disclosure enables more accurate modeling of a company’s business. This capabil-
ity minimizes inaccurate consensus modeling, which can lead to nasty earnings
surprises and huge disappointment-based sell-offs. It would be the height of
solipsism to assume that the parlay between management and Wall Street drives
the disclosure decision, though that doesn’t stop analysts from thinking it so.
       At the same time, companies are engaged in an ongoing balancing act
between providing too much data and too little. The nature of this kind of seg-
mented data can vary from CEO to CEO and from CFO to CFO, even within a
single company. Strategic realignment can also play a role. Consider the case of
General Electric, an acquisitive and dispositive company. Even without the pas-
sion for business development hardwired into GE’s DNA, the company has a
                                      Phase 2: Segment Modeling of Revenues •   65



penchant for revising its segment organization—and consequently its segment
presentation.
      When I covered the stock in the mid-1990s, GE’s Financial Service business
had five main competencies and 20 to 25 identifiable subsegments. In time, the
information flow lessened, complicated by Reg FD. Industrial GE, which once
reported about a dozen segments, began spinning off consumer units and devel-
oping new competencies (e.g., water and security). Simultaneously, GE was col-
lapsing its industrial businesses one into the other to reflect broad areas of
concentration rather than product segments.
      If an analyst covering GE from the 1980s or 1990s had stepped away and
only now resumed GE coverage, he or she would confront a nearly unrecogniz-
able segment representation. These continual changes represent a modeler’s
nightmare. The model as built must be flexible enough to accommodate change
as it occurs; and if you cover 12 to 20 companies, something is always
changing.


Revenue Modeling: Two Cases
Let’s consider the two major kinds of company presentations. The first is a com-
pany that provides segment revenue data, either on a product-line or geographic
basis. This company may also provide some degree of granularity related to a
complete or partial breakout of the individual revenue segments.
      The other kind of company provides segment revenue as well as some form
of segment income, be it operating income, EBIT, or pretax income. This level of
data permits not only a more granular revenue modeling process but also a way
to completely rethink the income statement model itself. We emphasize that this
unique modeling method, which we detail in the following paragraph, is an
option and not a requirement for successful modeling. But we also want to take
maximum advantage of information presented, and this is one way to do so.
      In the remainder of this chapter, we will discuss modeling up to the P&L
for companies that provide segment revenue only. In chapter 4, we will begin our
discussion with a treatment of companies that provide segment revenue as well
as individual segment income.


Revenue Data Only
The most common presentation format is revenue by segment, with little to no
below-the-line color and no segment operating income. All but the largest com-
66   •   Income Statement Presentation



panies can convincingly argue that segment margin disclosures provide too-
detailed information to competitors. There’s another interested party that many
companies prefer to keep in the dark regarding segment margins: their
customers.
       Smaller companies with a somewhat uniform product and geographic mix
in a segment in particular have a good reason for holding their segment operating
income data close to the vest. Simply, if a good customer buys goods or services
from company A on the underlying premise that this company is entitled to a
10% margin on the sale, and the segment in which this customer participates
reports a 5% operating margin, the aggrieved customer will conclude that other
customers are getting a better deal. Larger companies tend to aggregate so many
products and geographic markets in their segment presentation that the various
high or low margins accorded various classes of products get lost in the blend.
       There is a whole class of companies, some of them quite large, that once
provided segment revenue and operating income data but now only provide seg-
ment revenue. Some of these firms may have had concerns about providing too
much information to competitors and customers; others may have gotten tired
of publishing data about money-losing segments alongside data about profitable
business units.
       Ciena Corp. is a company that provides formal segment revenue data, which
is available in its SEC filings as well as its quarterly results release. The company
also provides less formal data points shared in other public forums, such as its post–
results release conference call and in various presentations that are archived on its
Web site. Let’s first concern ourselves with the representation and modeling of seg-
ment revenue that the company provides to everybody, including the SEC.
       The company began as a more or less pure play on optical communications
infrastructure. After years of business development mainly related to small-scale
acquisitions, Ciena has added various data networking, broadband access, storage
networking, and related capabilities to its core competency. In recognition that
its solution suite lacked various parts of an integrated network offering, the com-
pany has cast itself as a network specialist. The company’s segment presentation
has equally undergone various permutations.
       As of early 2009, Ciena reported three business units: Optical Service Deliv-
ery, Carrier Service Delivery, and Global Network Services. Ciena does not pro-
vide segment income information. Figure 3.1 shows Ciena’s revenue segments by
quarter for the October 2007 fiscal year. Pay attention to this illustration, because
it contains many of the basics in segment presentation and modeling. The seg-
ments are presented down the page and summed across the interim periods. In
                                         Phase 2: Segment Modeling of Revenues •   67



Figure 3.1 we compile sequential percentage-change comparisons and percentage
of revenue per segment. This percentage of revenue compilation will be useful
later on when, for example, we go more granular on gross margin.
       For our presentation of historical segment revenue, we again compile eight
quarters (four at absolute minimum) of historical segment data. We’ve already
created our 12-column annual presentation structure, so the historical data is
segmented logically beneath the respective quarters. We also sum for our rolling
summary periods, including half year, nine months, and full year. We populate
the adjacent annual percentage-change cells as well.
       To get the historicals on the page the way we like them, we reiterate our
earlier assumptions regarding presentation format: that modeled year 1 begins in
column CA, and eight quarters of historical data are to the left. Historical seg-
ment data for quarter 1 (two years prior) begins in cell BA70. In our title column
A, in A70 through A72, we list Optical Service Delivery, Carrier Service Delivery,
and Global Network Services. In column A73, we indent and type in “Total.”
       In our title column A, skip row 74. In cell A75, type in “Sequential & Percent-
age Revenue.” In A76 through A79, copy the three segments as well as Total.
       In our initial historical quarter cell, calculate Optical Service Delivery as a
percentage of total revenue; repeat for the two other segments. Leave the total
column blank, as this always sums to a percentage of 100 percent and provides
us no useful information.
       Let’s now use our annual comparison columns for some closer analysis. In
row 76, calculate a sequential comparison for Q1 of our beginning historical
period. We have no actual predecessor period in our model, but beginning in 2Q
these values become real as they do for subsequent 1Q periods. In our summary
columns, replicate percentage of revenue; we obviously have neither the place nor
method to model sequential change in the compilation periods. I would typically
italicize the entire sequential and percentage revenue presentation. This visually
differentiates it from the presentation of segment revenues immediately above.
       Once the quarterly data is entered for historical year two, we can copy and
paste the sequential and percentage revenue presentation for year 2; note that 1Q
now has a working sequential comparison. And we are ready to proceed to model
year 1.


Segment Revenue Modeling: Sequential or Annual?
For each revenue segment, the historical percentage growth on an annual basis
or sequential basis may stand as a sufficient modeling guideline. This is most
Figure 3.1

Ciena’s segment revenues for fiscal 2007. The presentation features both annual and sequential revenue percentage-change comparisons, along with percentage-
of-revenue contributions to the total top line.

 Ciena

 SEGMENTS                     1Q07        Y/Y       2Q07       Y/Y       1H07       3Q07       Y/Y    9mos07        4Q07       Y/Y       2007       Y/Y
 Optical Service Delivery       132.1        63%      159.5       68%      291.6       167.0      45%    458.6        180.9      46%       639.5       54%
 Carrier Service Delivery         9.4       -62%        9.6      -57%       19.0        15.2     -27%     34.2         12.8       -7%       47.0      -43%
 Global Network Services         18.8        30%       20.3       45%       39.1        22.8      55%     61.9         22.5      22%        84.4       37%
   Total                        160.3        33%      189.4       44%      349.7      205.0       36%    554.7        216.2      38%       770.9       38%

 Sqntl & Pcntg Rvnu           % Rvnu        Sqntl   % Rvnu       Sqntl   % Rvnu     % Rvnu       Sqntl   % Rvnu    % Rvnu        Sqntl   % Rvnu
 Optical Service Delivery        82%          7%       84%        21%       83%        81%         5%       83%       84%          8%       83%
 Carrier Service Delivery         6%        -31%        5%         2%        5%         7%        58%        6%        6%        -16%        6%
 Global Network Services         12%          2%       11%         8%       11%        11%        12%       11%       10%         -1%       11%
  Total Rvnu                                  3%                  18%                              8%                              5%
                                        Phase 2: Segment Modeling of Revenues •   69



likely to be the case for the casual modeler, or the buy-side generalist charged
with monitoring a broad basket of stocks.
      Even for those who will do a deeper dive on segment modeling in our Phase
3 worksheet, which is located in Chapter 4, for now we will base our current-year
and forward-year segment revenue modeling on historical growth rates. Don’t
forget we have a total five-year rear-view on the segment: the three annualized
segment totals (if available) as well as two years of individual historical segments
by quarter. We want to take care in analyzing the historical growth rates as a
starting point.
      There are various ways to “seasonalize” an assumed full-year revenue cal-
culation. But a year is a long, long time in the life of a corporation. We rarely
seasonalize something as large and lumpy as a full-year consolidated revenue
total. Instead, we are disposed to growing the revenue hypothetically from the
appropriate historical 90-day period, be it the immediately preceding one or one
from one year earlier.
      Which is the appropriate historical period? For heavily seasonal industries
such as retail, the best growth gauge is against the year-earlier period. For many
technology companies, the best gauge can be sequential growth. This can also be
true for resources-intensive industries such as energy and basic materials, where
the prevailing commodity price rather than the prior-year commodity price is
the primary top-line driver.
      Once a technology company gains consumer end-product exposure, how-
ever, the company’s formerly seasonless operating model may take on a seasonal
color—particularly around fourth-quarter holiday sales. For industrial compa-
nies and a whole host of other industries, various factors can introduce uncer-
tainty into the revenue modeling process; these include new product introductions,
competitive threats, economic cycles, and social and political events.
      The fact is that modeling segment revenues for most companies falls into
neither the purely sequential nor the purely annual camp. For companies with
neither a seasonless operating model nor a highly seasonal profile, our rule of
thumb is to use sequential growth if we have highly granular inputs such as indi-
vidual product family tallies at the subsector level. If our only guidance is very
general and at the segment level or even solely at the top-line level (customary
with retailers), we will model quarterly growth on a year-over-year basis.
      Your historical grid of segment revenues shows both annual and sequen-
tial growth rates for each segment. Depending on your industry, model
growth using the appropriate historical sequential or annual rate of change
for each period. Later, we’ll have a chance to go more granular at the segment
level as well.
70   •   Income Statement Presentation



Product and Services Revenue Within the P&L
A large number of companies introduce a highly useful level of granularity
directly into the income statement. These companies break out product revenues
from services revenues.
      We made the point earlier in this book that investors want the final data
input—the asset’s calculated value—and care not if you get the line items right
along the way. So, why pay such attention to the bifurcated revenue presentation?
Because it is typically accompanied by distinct COGS data, and that will give us
two inputs for our gross margin calculation. Along with operating margin, gross
margin is one of the most crucial determinants of the final bottom line.
      A weighted gross margin calculation for a company that provides product
and service data in the revenue and COGS lines is fairly straightforward; calcu-
late percentage of total revenue from products multiplied by its gross margin and
the percentage of total revenue from products multiplied by its historical gross
margin, and sum the two.
      Assuming a company does break out revenue by product and services, how
should the line items be modeled? There are several ways to model revenue up
from segments into this bifurcated top line, and as always the modeling method
begins with the company presentation. The simplest case (and less common than
you might think) is an instance where the presenting company that uses this
bifurcated top-line presentation also provides segment data that includes services
revenue, and this segment services revenue historically accords exactly with ser-
vices segment revenues.
      In this instance, one models services revenue at the segment line and plugs
the number into the P&L services revenue line item. For top-line product reve-
nue, one simply sums revenue contribution from all product-based segments.
      Doesn’t P&L services revenue always line up with segment services revenue?
Not always; and the logic is straightforward. The P&L is simply a reckoning sheet;
it divides consolidated revenues strictly by whether a dollar was recorded based
on the contribution from product sales or from services delivered. The segment
services, on the other hand, commonly represents a functioning business unit.
And the service personnel in this unit, in the course of their daily tasks, acquire
product at arm’s length from the product segments, which is then installed and
serviced as needed. Intuitively in this instance, the segment services revenue will
be higher than P&L services revenue.
      Figure 3.2 shows ADC Telecom’s fiscal 2007 top line, including the product-
to-service breakout, as well as the segment breakout. For ADC Telecom’s October
2006 through 2008 fiscal years, P&L services revenue ranged between 74% and
Figure 3.2

ADCT’s fiscal 2007 P&L services revenues do not line up with its segment services revenues because the Global Services unit acquires product internally that it
then resells in the course of doing business, whereas the top line reflects only revenue earned purely from services.

ADC Telecom
Income Statement                  1Q07 Yr/Yr %          2Q07 Yr/Yr %         1H07      3Q07 Yr/Yr % 9mos07             4Q07 Yr/Yr %         2007     Yr/Yr %
Prodct Sales                       259.64    7%          312.00   -3%          571.6    306.00   -0%   877.6            290.50    9%         1,168.1       3%
Service Revenue                     36.85   -7%           37.40    0%           74.3     40.10  11%    114.4             39.10   -6%           153.5      -1%
    Net Sales                      296.49    5%           349.4   -2%          645.9     346.1    1%  992.0              329.6    7%         1,321.6       2%
Product COGS                       168.80    7%          195.90   -6%          364.7    196.80   -2%   561.5            182.50    1%         744.00       -0%
Service COGS                        33.40   -8%           33.00  14%            66.4     35.70  19%    102.1             33.50   -2%         135.60        5%
    Net COGS                        202.2    4%           228.9   -3%          431.1     232.5    0%  663.6              216.0    1%           879.6       0%
 Gross Profit                         94.3    8%           120.5   -1%          214.8     113.6    1%  328.4              113.6  22%           442.0        7%

Segments                             1Q07                 2Q07                 1H07       3Q07              9mos07       4Q07                  2007
Revenues
 Global Connectivity                 228.8       10%      272.1       -3%      500.9      268.7       -1%      769.6      251.4      10%     1,021.0       3%
 Network Solutions                    22.5                 26.5                 47.3       24.5                 71.8       24.4                 96.3
 Professional Services                47.6       7%        50.8       -3%       98.4       52.9       4%       151.3       53.7       3%       205.0
 Total Revenues                      297.2       9%       349.4       -2%      646.6      346.1       1%       992.7      329.6       7%     1,322.3       1%
 PrSrvc % of P&L Srvcs                                                                                                                          75%

 Segment Srvcs as % P&L Srvcs                                                                                                                  75%
72   •   Income Statement Presentation



77% of segment services revenue. To model this relationship up from the seg-
ments to the top line, we would represent P&L services revenue as 75.5% of esti-
mated segment services revenue. Top-line product revenue would be the sum of
the product segments, plus segment services revenue minus P&L services reve-
nue. (As a sign of how company presentations can change, by fiscal 2008 ADCT’s
P&L services revenue was lining up with its services segment revenue; and by
2009 ADCT was no longer showing a services component of revenue on the P&L
within its quarterly results releases.)
      Some companies that provide top-line breakouts of services versus products
do not break out a distinct services revenue component within the segment
reporting. To model services revenue for these companies, we’ll initially rely on
the historical amount of services revenue as a percentage of total revenue.
      Now that we have summed segment revenue in line 73 of our example, it is
a simple step to link this to total revenue in the top line for each quarter. Com-
panies may show all in top-line revenue (product and service revenue combined);
they may break out product and services revenue in which P&L services revenue
aligns exactly with segment services revenue; or they may present product and
services revenue in which services revenue is out of sync with (and usually lower
than) segment services revenue.
      If the company has only one top-line revenue number, link the summed
segment revenue into that cell. If the company has an exact relationship between
P&L services revenue and segment services revenue, link the summed segment
tally directly into top-line services revenue. In the P&L product revenue cell, link
to the sum of all product-based divisions.
      If, like ADCT, this relationship between P&L and segment services revenue
is inexact, we need an adjustment. Assuming P&L services revenue is in CA4 and
segment services revenue is in CA72, our formula in cell CA4 would be,
   CA72*.75. For P&L product revenue, our formula in CA3 would be
   CA70 CA71 (CA72*.75). A final step is to establish these links for all quar-
ters. We’ve already summed the “placeholder” values in each quarter across the
interim periods, so by linking summed segment revenue to all top-line cells, all
our interim-period tallies (including full year) will be in place.
      We’ve shared some of the common methods by which individual companies
present their segment data, and how to use those common presentations to more
accurately model the top line. While most presentations hew to the methods we’ve
discussed, you will encounter other presentation modes in your analytic career.
      Further along we’ll discuss other means to refine the inputs within the
individual segment line items. But first we offer a unique method to reimagine
the income statement presentation altogether, in which segment inputs take pre-
eminence over the normal operating cost line items.
                                                  Chapter 4
           PHASE 3:
           SEGMENT OPERATING
           INCOME AND
           PERCENTAGE-OF-
           DIFFERENCE
           MODELING


“Percentage of Difference”—Modeling the P&L from
Segment Income
For companies that provide both segment revenue and segment operating income
(or EBIT), we model segment revenue as well as segment income. Later in this
chapter, we show how to use modeled segment income directly to recast the
income statement.
      Even if that option is not available or not preferred, we model this level of
detail. Additionally, we display the margins for each segment for each quarter and
interim periods below the operating profit bloc.
      It is not uncommon for analysts to use the historical margin percentages,
along with company guidance and their own analysis of industry dynamics, to
forecast segment margins and—applying those margins to estimated segment
revenues—to forecast segment operating income. The sum of segment operating
profits, adjusted by the historical percentage eliminations, should be near the
analyst’s forecast of operating income modeled in the P&L. Analysts will use this
information to calibrate their P&L operating income.
      While it is useful to “eyeball” estimated segment operating profits in rela-
tion to estimated P&L operating profits, whenever possible we want to quantify

                                                                                73
74   •   Income Statement Presentation



such relationships. Hence, we use what we call percentage-of-difference modeling
to make these tallies not just close but identical.


Percentage of Difference: The Concept
Percentage-of-revenue representation is a mainstay of P&L modeling. When done
right, it provides precision—and even some satisfaction, most of it internal to the
modeler. If you’ve modeled a company’s quarter on a percentage of basis and that
company reports line items that are spot on with, say, your SG&A and R&D
assumptions, be sure to give yourself a high five—because no one else cares. We’ll
say it again, and frequently, in this book: the important outputs are conclusions,
not components.
       If a company provides segment income data along with segment revenue,
there are special instances in which this information can be used to model the
P&L. Rather than flowing down strictly from the top, this kind of modeling
emanates both up and down from the middle and specifically from operating
income modeled at the segment level. The “difference” in percentage of differ-
ence is that between revenue and operating income.
       For this kind of modeling, we require a company that regularly provides
investors—via its financial press releases and/or its 10-Q—with both segment
revenues and segment income. Moreover, at some place in the income statement
presentation, we require a point at which P&L operating income accords exactly
with segment operating income. One such company is Motorola; its P&L operat-
ing income accords exactly with the sum of its segment income (including the
other and eliminations line). During 2006, we modeled Motorola using percent-
age-of-difference modeling. Hence, the segment operating profit tallies (which
were given their own line item on row 11 in the P&L) lined up exactly with the
segment earnings, and with the P&L operating income estimate. Figure 4.1
shows the historical presentation of Motorola’s 2006 year; we’ve included a snip-
pet of the P&L up to operating income, as well as segment revenue and operating
income, to show how P&L and segment operating income line up exactly.
       At the segment earnings level, some companies may present EBT (earnings
before taxes); one such company is Qualcomm. For Qualcomm, the sum of seg-
ment income along with what it calls corporate and eliminations matches exactly
its pretax income; in other words, nonoperating items such as interest income
have been allocated to the segments. The corporate and eliminations figure for
Qualcomm was historically always positive, because it included the sizable Trea-
sury contribution of interest income. Now that Qualcomm like other companies
has had to restructure, this line item is sometimes negative.
Figure 4.1

Using percentage-of-difference modeling, modeled segment operating income and P&L operating income become identical. Pictured is data from Motorola’s
2006 year.

 Motorola Inc.
                                 1Q06                 2Q06                  1H06       3Q06                 9mos06     4Q06                 2006         Y/Y %
 Revenue                          10,013     23%       10,876      23%       20,889     10,603      13%       31,492    11,792      13%      42,879         16%
 COGS                              6,993     27%        7,517      26%        14,510     7,226      13%       21,736     8,723      22%      30,152         21%
   Gross Margin                    3,020     13%        3,359      17%         6,379     3,377      12%        9,756     3,069      -7%      12,727          7%
 SG&A                               1,116    11%         1,195     16%         2,311      1,174     15%        3,485      1,158      1%       4,504          7%
 Restructuring & Other Charges          -                    -                                -                              65                 133
 R&D                                 980      21%        1,016      19%       1,996      1,027      20%        3,023     1,061      16%       4,106          20%
 Other/FreeScl                         30   -529%       (374)    -2178%       (344)        205                 (139)         32   -106%       (108)
   P&L Op Incm                       849                1,522       55%       2,416        968     -12%        3,387        753    -57%       4,092         -13%
   Segment Op Incm                   849                1,522                 2,371        968                 3,339        753               4,092
   Pro Forma Op Incm                 976                 1,196                 2,172     1,220                 3,392       894                4,286

 Segment Sales                     1Q06                 2Q06                  1H06       3Q06               9mos06       4Q06                  2006
 Mobile Devices                    6,403    45.0%        7,140    45.7%      13,543      7,034     25.5%     20,577       7,806    19.4%     28,383         32.3%
 Home & Networks                   2,520     -7.9%      2,343     -17.1%      4,863      2,262    -18.2%       7,125      3,043     5.9%      10,168        -9.2%
 Enterprise Mobility                 732      7.3%      1,355     82.4%       2,087      1,329     78.9%       3,416        980    39.4%       4,396        53.1%
 Other                                 -          -          -          -         -          -          -          -          -         -          -             -
 Eliminations                       (47)      0.0%        (18)   -71.0%        (65)       (22)    -66.2%        (87)       (37)   -51.3%       (124)       -54.1%
 TOTAL                             9,608    22.6%      10,820     28.7%      20,428     10,603     17.2%      31,031     11,792    17.5%     42,823         21.4%

                                                                                                                                                       (continued)
Figure 4.1 (continued)


 Motorola Inc.
                         1Q06                2Q06                1H06      3Q06                 9mos06     4Q06                 2006      Y/Y %
 % Rvnu & Sqntl          %Sales    Q/Q %     %Sales    Q/Q %     %Sales    %Sales    Q/Q %       %Sales    %Sales    Q/Q %      %Sales
 Mobile Devices           66.6%      -2.1%    66.0%     11.5%     66.3%     66.3%      -1.5%      66.3%     66.2%     11.0%      66.3%
 Home & Networks          26.2%    -12.3%      21.7%     -7.0%    23.8%      21.3%    -3.5%       23.0%     25.8%     34.5%      23.7%
 Enterprise Mobility        7.6%      4.1%    12.5%     85.1%     10.2%     12.5%      -1.9%      11.0%       8.3%   -26.3%      10.3%
 Other                      0.0%      0.0%      0.0%      0.0%      0.0%      0.0%      0.0%       0.0%       0.0%     0.0%        0.0%
 Eliminations              -0.5%   -38.2%      -0.2%   -61.7%      -0.3%     -0.2%    22.2%       -0.3%      -0.3%    68.2%       -0.3%
 TOTAL                              -4.3%               12.6%                         -2.0%                           11.2%

 PF Operating Profit
 Mobile Devices              702     60.3%       804     63.1%     1,506      843     42.2%        2,349      341    -48.6%       2,690    23.0%
 Home & Networks             161    -61.2%       222    -55.1%       383       181    -61.1%         564      428     -21.0%        992   -48.2%
 Enterprise Mobility         142                 239   624.2%        381      254    551.3%          635      118    126.9%         753   466.2%
 Othr & Eliminations       (156)   -351.6%       257      0.0%       101     (310)          -      (209)        -           -     (209)
  Segment Totals             849     -8.1%     1,522     58.9%     2,371      968    -10.4%        3,339      753    -56.0%       4,092    -11.1%

 Operating Margin
 Mobile Devices           11.0%               11.3%               11.1%     12.0%                 11.4%      4.4%                 9.5%
 Home & Networks           6.4%                9.5%                7.9%      8.0%                  7.9%     14.1%                 9.8%
 Enterprise Mobility      19.4%               17.6%               18.3%     19.1%                 18.6%     12.0%                17.1%
 Eliminations              0.0%                0.0%                0.0%      0.0%                  0.0%    362.2%               108.1%
 Segment Margin            8.8%               14.1%               11.6%      9.1%                 10.8%      6.4%                 9.6%
       Phase 3: Segment Operating Income and Percentage-of-Difference Modeling   •   77



     At first blush, the distinction between Motorola’s presentation and that of
Qualcomm may seem insignificant. But it is infinitely simpler to model up to
P&L operating income from segment operating income than it is to model up to
P&L EBT from segment EBT. The wild card in EBT is the uncertainty around any
quarter’s net interest cost/income.


Percentage of Difference: Basis
As always, a walk-through best makes our case. We’ll begin doing our walk-
through with Motorola, and we’ll first consider a period in which Motorola was
profitable. Then we’ll up the ante by attempting to model during one of the many
unprofitable phases of this “serial” restructurer.
      With that in mind, in Figure 4.2 we take another look at the alignment
between P&L profit and summed segment profit at Motorola. During 2006,
everything was working for Motorola; in the third quarter, as sales of its ultraslim
RAZR mobile phone took off, margins hit their peak. All three of its divisions
(actually, these were historically recast from a preceding organizational struc-
ture) were profitable. In 2Q06, Mobile Devices had an 11.3% operating margin;
Home and Networks had a 9.5% margin; and Enterprise Mobility had a rich
17.6% margin. We use these hypothetical margins to model segment income,
adjusting for some volume leverage for this pre-Christmas period in which sales
were rising as dealers stocked inventories (as it turned out, 3Q06 segment mar-
gins were higher for two of the three divisions).
      In 2Q, as opposed to other periods, other and eliminations contributed
rather than deducted from segment operating income. For all of 2006, other and
eliminations at the segment level was a cost of $209 million, on total operating
income of $4,100. This item typically reduces operating income by 5%. In mid-
year 2006, having recorded the 2Q actual, we would have tweaked our 3Q seg-
ment income and margins assumptions. Our segment income calculation for
3Q06 at the time was the respective segment margins times forecast revenue and
summed, with the sum reduced by 5% to reflect other and eliminations. On a
modeled basis, our summed and adjusted segment operating income—which
anticipates a cost rather than a gain in others and eliminations—is 9.1% of rev-
enue, or $968 million.
      We need a line in our P&L model to accommodate this calculation. So,
between GAAP operating income and pro forma operating income in the P&L
we insert a line to accommodate what we call segment operating income. In
that line, we pull up our segment income calculation from the segment
model.
Figure 4.2

Another look at Motorola from the perspective of percentage-of-difference modeling.

 Motorola Inc.
                                  1Q06                  2Q06                  1H06       3Q06                 9mos06     4Q06                 2006       Y/Y %
 Revenue                           10,013       23%      10,876      23%       20,889     10,603      13%       31,492    11,792      13%      42,879       16%
 COGS                               6,993       27%       7,517      26%        14,510     7,226      13%       21,736     8,723      22%      30,152       21%
   Gross Margin                     3,020       13%       3,359      17%         6,379     3,377      12%        9,756     3,069      -7%      12,727        7%
 SG&A                                1,116      11%        1,195     16%         2,311      1,174     15%        3,485      1,158      1%       4,504        7%
 Restructuring & Other Charges           -                     -                                -                              65                 133
 R&D                                  980       21%        1,016      19%       1,996      1,027      20%        3,023     1,061      16%       4,106       20%
 Other/FreeScl                          30    -529%       (374)    -2178%       (344)        205                 (139)         32   -106%       (108)
   P&L Op Incm                        849                 1,522       55%       2,416        968     -12%        3,387        753    -57%       4,092      -13%
   Segment Op Incm                    849                 1,522                 2,371        968                 3,339        753               4,092
   Pro Forma Op Incm                  976                  1,196                 2,172     1,220                 3,392       894                4,286

 Segment Sales                      1Q06                  2Q06                  1H06       3Q06               9mos06       4Q06                  2006
 Mobile Devices                     6,403     45.0%        7,140    45.7%      13,543      7,034     25.5%     20,577       7,806    19.4%     28,383      32.3%
 Home & Networks                    2,520      -7.9%      2,343     -17.1%      4,863      2,262    -18.2%       7,125      3,043     5.9%      10,168     -9.2%
 Enterprise Mobility                  732       7.3%      1,355     82.4%       2,087      1,329     78.9%       3,416        980    39.4%       4,396     53.1%
 Other                                  -           -          -          -         -          -          -          -          -         -          -          -
 Eliminations                        (47)       0.0%        (18)   -71.0%        (65)       (22)    -66.2%        (87)       (37)   -51.3%       (124)    -54.1%
 TOTAL                              9,608     22.6%      10,820     28.7%      20,428     10,603     17.2%      31,031     11,792    17.5%     42,823      21.4%
     Figure 4.2 (continued)


      Motorola Inc.
                              1Q06                2Q06                1H06      3Q06                 9mos06     4Q06                 2006      Y/Y %
      % Rvnu & Sqntl          %Sales    Q/Q %     %Sales    Q/Q %     %Sales    %Sales    Q/Q %       %Sales    %Sales    Q/Q %      %Sales
      Mobile Devices           66.6%      -2.1%    66.0%     11.5%     66.3%     66.3%      -1.5%      66.3%     66.2%     11.0%      66.3%
      Home & Networks          26.2%    -12.3%      21.7%     -7.0%    23.8%      21.3%    -3.5%       23.0%     25.8%     34.5%      23.7%
      Enterprise Mobility        7.6%      4.1%    12.5%     85.1%     10.2%     12.5%      -1.9%      11.0%       8.3%   -26.3%      10.3%
      Other                      0.0%      0.0%      0.0%      0.0%      0.0%      0.0%      0.0%       0.0%       0.0%     0.0%        0.0%
      Eliminations              -0.5%   -38.2%      -0.2%   -61.7%      -0.3%     -0.2%    22.2%       -0.3%      -0.3%    68.2%       -0.3%
      TOTAL                              -4.3%               12.6%                         -2.0%                           11.2%

      PF Operating Profit
      Mobile Devices              702    60.3%        804     63.1%     1,506      843     42.2%        2,349      341    -48.6%       2,690    23.0%
      Home & Networks             161    -61.2%       222    -55.1%       383       181    -61.1%         564      428     -21.0%        992   -48.2%
      Enterprise Mobility         142                 239   624.2%        381      254    551.3%          635      118    126.9%         753   466.2%
      Othr & Eliminations       (156)   -351.6%       257      0.0%       101     (310)          -      (209)        -           -     (209)
       Segment Totals             849     -8.1%     1,522     58.9%     2,371      968    -10.4%        3,339      753    -56.0%       4,092    -11.1%

      Operating Margin
      Mobile Devices           11.0%               11.3%               11.1%     12.0%                 11.4%      4.4%                 9.5%
      Home & Networks           6.4%                9.5%                7.9%      8.0%                  7.9%     14.1%                 9.8%
      Enterprise Mobility      19.4%               17.6%               18.3%     19.1%                 18.6%     12.0%                17.1%
      Eliminations              0.0%                0.0%                0.0%      0.0%                  0.0%    362.2%               108.1%
      Segment Margin            8.8%               14.1%               11.6%      9.1%                 10.8%      6.4%                 9.6%
79
80   •   Income Statement Presentation



Percentage of Difference: Execution
A key takeaway is that the percentages we will use must sum to 100% of revenue.
Normally, in a pure P&L-based model, our modeling of various percents of rev-
enue for the main line items would sum to the difference between 100% and the
operating margin. Using these estimates, we would arrive at costs as a percentage
of revenue of, say, 93% for an industrial company or 73% for a software company.
In percentage-of-difference modeling, we’ll estimate our operating margin at the
segment level and then let that margin level drive the manufacturing and operat-
ing cost inputs, rather than the other way around.
      In modeling Motorola’s 3Q06 at the P&L level, consolidated revenue as
always comes up from the summed segment revenue. But in this case we model
cost of goods sold not as a percentage of revenue but as a percentage of the differ-
ence between revenue and summed segment operating income of 9.1%. The other
main line items of operating cost—SG&A and R&D—will also be represented
not as a percentage of revenue but as a percentage of the difference between sales
and our modeled 9.1% margin.
      All of our line items have to reflect that 9.1% of revenues that has been
captured as operating income. We would adjust for this margin by increasing the
line items by this amount. With Motorola’s COGS running at about 68% at the
time, our COGS formula would begin with 68% 1.091 = 74.4%. In a scenario
in which 3Q06 revenue is cell CV3, COGS is cell CV4, and segment operating
income is cell CV16, the formula for COGS will be (CV3 CV16)*.74.4.
      SGA has been a bit over 11%, and R&D has been 9.7%. These items must
be similarly “squared up” by that 1.091 factor. We also know Motorola’s catch-all
“other” category typically runs at 2% of revenue; it too must be adjusted. Adjusted
by this 1.091 multiplier or factor, SG&A is 11.05% times 1.091, or 12.1%; R&D is
9.7% times 1.091, or 10.6%; and Other is 2.0% times 1.091, or 2.2%.
      In a scenario in which 3Q06 SG&A is cell CV6, our SG&A formula is
(CV3 CV16)*.12.1. In cell CV8, our R&D formula is (CV3 CV16)*.106; and in
cell CV9, our other formula is (CV3 CV16)*.022.
      The four multipliers sum to 99.3, so we need to bump them up to sum to
100%. The multiplier for COGS becomes 74.5; the multiplier for SG&A becomes
12.4; the multiplier for R&D becomes 10.8; and the multiplier for “other” becomes
2.3. These sum to 100%. How can we be so cavalier about adjusting these per-
centage-of-revenue tallies? Because, remember, no one is paying you for accuracy
in calculating COGS, R&D, or SG&A; these calculations are just means to an end
in modeling.
       Phase 3: Segment Operating Income and Percentage-of-Difference Modeling   •   81



     At this point, you will see that the segment income line item is exactly
equivalent to the GAAP P&L calculation of operating income. Then, news flash:
the company preannounces that, due to a shortage of capacitors in Kazakhstan,
phone margins will suffer. Based on your determination that this event will cost
100 basis points of Mobile Device segment margin, you shave your 3Q06 margin
assumption for mobile devices to 11% from 12%. The dollar value of your seg-
ment operating income is reduced. And a funny thing happens: so too is your
P&L operating income assumption, by exactly the same amount.


Percentage of Difference: From GAAP to Pro Forma
As you know, we’ve made a determination to include line items in our income
presentation to allow for pro forma modeling, either because this is de rigueur for
the company or because the company has entered into a rough period in which
restructuring and impairment costs must be accommodated without distorting
go-forward progress.
      For the company for which pro forma operating income is the exception
rather than the rule, we can simply equate summed segment operating income
to GAAP P&L operating income as well as pro forma P&L operating income. We
can then accommodate those special cases when the company must incur costs
to restructure or impair.
      As we know, many companies always exclude noncash items (intangibles
amortization, FAS 123R stock option compensation, acquired in-process R&D)
and routinely differentiate pro forma operating income from GAAP operating
income. For these companies, at the segment level we model while excluding
these items from segment income.
      We’re aided by the fact that most companies that provide segment revenue
and segment income will provide adjusted as well as GAAP segment income
(adjusted to exclude one-time events and/or noncash items). Armed with histori-
cal comparisons, we can conclude how the inclusion and exclusion of such items
will impact individual segment margins.
      Now that we’ve determined which kind of company we’re dealing with—
serial excluder, occasional excluder, or never-say-die includer—we can direct the
summed segment operating income to the appropriate cell. If the summed seg-
ment income is modeled to exclude noncash items and anticipated one-time
restructuring or impairment costs, the segment operating income cell is modeled
to match up or exactly reconcile to the cell for pro forma P&L operating income.
If the summed segment income is modeled to include noncash items and antici-
82   •   Income Statement Presentation



pated one-time restructuring or impairment costs, the segment operating income
cell is modeled to match up or exactly reconcile to the cell for GAAP P&L operat-
ing income.
       From this point, we then proceed down the P&L presentation to adjust for
all the nonoperating items on the way to our GAAP and pro forma net income
and per-share calculations.
       Modeling segment revenue and income will be sufficient for many investors
and analysts, particularly those charged with covering a large group of companies
or industries. But it is possible and even preferable to model each revenue seg-
ment at a more detailed level. The difficulty is that each revenue segment is sub-
ject to multiple forces; modeling at this granular level is not easily categorized or
systematized. In our next chapter, we offer some guidance on this broad-based
topic.
                                                  Chapter 5
           PHASE 4:
           THE WORKBENCH,
           PART 1




Down below the income statement, margin analysis, and the segment model in
every workbook-based model we build is an area variously called the workbench,
the worksheet, the work zone, or some similar such construction-related term.
No two of these workbench presentations are ever the same, even for companies
that organize their reported information in similar ways. The worksheet is almost
always a means of deriving a more granular revenue assessment, which in turn
improves the efficacy of the segment revenue model. The workbench can also be
directly and indirectly helpful in shaping our segment margin assumptions. And
only when we are satisfied with our revenue model do we fine-tune our gross
margin and operating cost assumptions
      Data from the workbench is used to model individual segment revenue.
Much as the summed segment revenue is linked to the cell representing consoli-
dated company revenue, the summed worksheet data per segment is directed to
the cell representing individual segment revenue.
      In terms of income statement presentation, representation, and modeling,
this will be by far our most (and in fact only) touchy-feely chapter. Beyond tran-
scribing information from SEC documents, analysis is not a quantifiable science,
the groaning shelves of analysis treatises notwithstanding. It is a mix of science
and art, with a subtext of detection—which necessarily involves a fair amount of
gumshoeing.
                                                                               83
84   •   Income Statement Presentation



       The workbench straddles publicly available, company-furnished company
information that is not part of the formal segment model; anecdotal and oral
company information; industry information; economic information; analyst
gossip; and all matter of directly or indirectly pertinent blog blather, scuttlebutt,
rumor, and nonsense. It is up to the gumshoe-analyst to determine and incorpo-
rate only those data points that reliably and predictably enhance the quality of
the segment modeling.
       So far in our discussion, we have mainly made use of common modeling
techniques (e.g., percentage-of-revenue modeling) and less common techniques
(e.g., percentage-of-difference modeling up from segment income). Always, we’ve
worked with a fairly straightforward information set, which has informed a fairly
straightforward set of assumptions.
       The workbench is a different animal. It incorporates informal as well as
formal data. Much of the best and easiest-to-use information is available for sale;
like any scarce good, it is not cheap.


Culling Information Sources for the Workbench
The workbench is the place where the good analyst exploits whatever applicable
information is afloat in the ether—and I use that term pointedly. The Internet is
a great source of secondary information about a company. The first place to start
is with trade journals. Thanks to the Internet, this species—endangered by paper,
printing, and shipping costs—has revived and is indeed flourishing. Because
advertising by industry participants supports electronic publishing and operating
costs, most trade journals are free. And while most require some industry affili-
ation to qualify for a subscription, most journals consider financial services an
applicable and acceptable affiliation.
      Over the years, trade journals have often developed data tables that industry
insiders regard as indispensible. The creative analyst can make excellent use of
this data. At various times in my career, I covered the automotive industry, aero-
space/defense, and the chemicals industry. Chemical Week carried extensive
hydrocarbon, monomer, and polymer pricing tables down to the product level,
both on a merchant basis and a spot basis. Within the worksheet sections of vari-
ous models for companies such as Dow, DuPont, and Lyondell, I used changes in
the pricing data to accurately forecast changes in various business units sensitive
to pricing trends.
      Automotive Week regularly publishes monthly, quarterly, and year-to-date
data not only on individual companies and brands but also on per-vehicle unit
                                                Phase 4: The Workbench, Part 1 •   85



shipment data. Coupled with wholesale pricing data and assumptions culled
from the Internet, which was a much smaller and more manageable beast in the
late 1990s, I was able to fairly accurately project revenues for the various units
and whole companies. Using airplane shipment data from Boeing as well as
industry wholesale pricing information picked up in Aviation Week, I was able to
model revenue for Boeing.
      As the above implies, don’t look for much creativity in trade journal titles.
But do look to these publications for invaluable industry insight along with the
data. It is also true that you can purchase much more comprehensive industry
data that dissects every data point by user, geography, customer type, and more.
Such data is expensive. For the multi-industry buy-side analyst charged with
monitoring 10 industries within a few sectors, the cost would be prohibitive.
      Beyond trade journals, simple Internet searches can direct you to industry
blogs, user groups, even disgruntled former employees with an axe to grind.
Amid the relentless boosterism provided by the spin doctors in Investor Rela-
tions, (IR machine), such axes can cut away the clutter and provide rare insider
insight.


Company Information
Often the best source of information about a company is the company itself, but
the interchange between analyst and management is rarely a full and frank
exchange of views. The frequent forum for this exchange is the post–results
release investor conference call. During these calls, the ability to ask questions is
often limited to industry professionals, meaning investment analysts from major
and midsized brokerages. In this setting, the interplay between company officials
and analysts falls somewhere between a poker game and a minuet.
      Companies, as we have noted, must forever balance their information dis-
closure; they cannot afford to release too much detail or it might tip off rivals
or, worse, aggravate their major customers. At the same time, managements
want to highlight areas of progress and successful execution. Often, particu-
larly in down markets or amid withering competition, these highlights can be
hard to find.
      Analysts must never overlook the fact that all company presentations,
regardless of their veneer of cool objectivity, pass through the hand of skilled spin
doctors. In this spirit, a company’s quarterly presentation may trumpet an
obscure niche (e.g., “coat hanger sales in Burma tripled”) while conveniently
overlooking inconvenient facts: as coat hanger prices in Burma went from $10 to
$30, total coat hanger sales fell $100 million.
86   •   Income Statement Presentation



     For all that, companies do provide useful and what I would call anecdotal
information—anecdotal in the sense that it is spoken from prepared or ad lib
remarks, but not published in any formal documents filed with the SEC or even
on the company’s own Web site.


Impediments to Company Information
In the months and years after the market collapse of 2000–2002, regulators sought
a variety of remedies for the real and perceived ills that had at the least exacerbated
if not triggered the market’s downward skid. Demonstratively closing the barn door
well after the value had vanished, regulators came up with a few fixes, including
Sarbanes-Oxley and Reg FD. Sarbanes-Oxley, enacted in 2002 in the wake of Enron,
Tyco, and other scandals, sets a higher standard for public companies and the
accounting firms that certify their financial statements. For investors, however, the
well-intentioned act is viewed with some skepticism. On the upside, it requires top
executives to sign off on the veracity of SEC documents; on the downside, it is a
cost and time burden for smaller public companies.
       Reg FD, or Regulation Fair Disclosure, is noble in intent and flawed in
execution. Reg FD mandates that no employee of a public company may share
information with one or more investors that is not available to the public. It was
meant to eliminate the information advantage that bulge bracket analysts gar-
nered from cozying up to the CFO or CEO—bits of insider information casually
discussed over golf or at the 19th hole.
       In the Reg FD era, IR officials and CFOs are well schooled in information
distribution. IR personnel won’t go much off the official results release script
during follow-up calls. Inadvertent admissions on a morning conference call
often prompt an afternoon press release. But the most privileged information,
oddly, is available to one source: industry analysis firms. These firms gather
information in confidence from public firms, slice and dice it for useful con-
sumption, and put it up for sale.
       For technology companies, multiple information providers will slice and
dice sales data from every niche and every region in every imaginable way. For
energy companies, there are multiple firms providing reserve, exploration, refin-
ing, and pricing data for every product from crude through the countless
distillates.
       This information, thus collected, collated, and pinned up for inspection like
so many butterflies, is then purchased by those same cash-rich bulge bracket
firms that used to get their information anecdotally on the golf course but now
get it neatly organized in tables and charts. The individual investor, the buy-side
                                                Phase 4: The Workbench, Part 1 •   87



multi-industry generalist, the private money manager—all are effectively priced
out of this information grid. Meant to level the playing field, Reg FD ultimately
results in even more exclusive information segregation than the system it
supplanted.


Formal Company Presentations
Investor interactions in such public forums as post–results release conference
calls and corporate presentations can be useful in the modeling process, both for
specific periods and in shaping the modeling process overall. The two parts of
any such company-investor interaction have a completely different feel. The for-
mal presentation is tightly scripted and controlled. The Q&A follow-up can be
much more freewheeling.
       The formal portion of industry presentations and post–results release calls
can be highly informational. These meetings are carried over the Web and thus
meet the threshold of public availability. Accordingly, company officers will reg-
ularly disclose data that is never published in SEC documents or on the compa-
ny’s Web site, which only lives on in transcripts of these meeting.
       When Juniper Networks was a smaller company, the CEO at the time began
every discussion by listing the unit shipment numbers for router chasses and
blades (i.e., discrete devices that fit into chassis slots.) This information existed
nowhere else on the company’s Web site, in its SEC documents, or even in the flip
books used for corporate presentations. The chassis and port unit data was
immensely helpful in analyzing and modeling unit and pricing trends in Juni-
per’s infrastructure business. Like all good things, it was doomed to go away; now
that Juniper is a bigger company, it has. The replacement data is less directly
informative, but it too can be used to get under the covers of the published
information.
       In another example, Motorola accompanies its quarterly results release with
a sizable PDF presentation published on the company’s Web site. At one time,
that presentation included handset unit data segmented by standard or protocol
(e.g., GSM, iDEN, or CDMA 1x). In time, Motorola changed its presentation to
one in which it presented its units as percentages shipped to various regional
markets. Both data series were and are highly useful, and you will search high
and low without finding them anywhere in Motorola’s official publications.
       No one in the analytic community is happy to see any such anecdotal or
oral data series discontinued. The data is useful in the absolute, but most useful
in the way it informs quarter-over-quarter changes within the series, which in
turn correlates to broad segment trends.
88   •   Income Statement Presentation



Post–Results Release Q&A
After a company offers its formal remarks in such settings, the floor is opened to
questions. We count on analysts to be unbiased and analytical, but representa-
tional heuristics can loom large. (Translation: the osmotic gain in self-worth
from time spent with a high-powered executive can subconsciously sabotage our
objectivity.) Representational heuristics means judging a person based on his or
her relationship to a mental category. Still, most analysts can put aside their col-
legial feelings toward corporate officers long enough to ask hard questions.
      The surface civility of the Q&A session may seem like an extension of the
formal remarks presentation, but beneath the surface, tension bubbles. Analysts
are dedicated to prying loose useful information; IR people and C-levels (i.e.,
chief executive officer, chief financial officer, and other top executives) want to
divulge nothing more than is in the script. If you are an analyst who covers 15 to
25 companies and have been in the business for 10 years, you will have listened
to (or read the transcript of) at least 1,000 such calls. If you are not by then wise
to the nuances of such interactions, you’re in the wrong business.
      Specifically, analysts seek to pry apart and amplify the available informa-
tion, such as line items and margin data on the P&L, segment data on a product-
line or regional basis. They tug at the threads of broad announcements (e.g.,
partnerships, asset sales, restructuring, and impairments) in hopes of unraveling
corporate strategy. Figure 5.1 shows a truncated version of Ciena’s income state-
ment presentation, including workbench. Information on the main revenue driv-
ers comes from the post-results conference call and is rarely published elsewhere;
information on the minor revenue contributors is an estimate and comes from
discussions with management.
      Having been blunted in direct questioning, analysts seek not the head-on
approach but the off-directional approach; rather than seek actual undisclosed
and guarded customer, product, or regional data, they seek clues on the dynamic
or direction—the trend—of such subsegment data. Some analysts always pick at
a particular vein. One communication industry analyst always begins his ques-
tions with a request for gross margin information on a segment basis. That’s a
fairy shrewd opener for any analyst working in an industry with lots of segment
revenue data but limited segment profit data. Operating costs are more apt to be
shared in an efficient organization, so gross margin differences are the primary
profit drivers for individual segments. While this analyst rarely gets a direct
answer to his gross margin questions, he is provided sufficient directional data
to better model EBIT for those segments at companies where profit is not
disclosed.
Figure 5.1

In this compacted view of Ciena’s income statement presentation, we see the segment and subsegment modeling used to get to the P&L top line.

 Ciena
 Income Statement                 1Q09       Y/Y %     2Q09      Y/Y %     1H09       3Q09        Y/Y %  9mos09       4Q09E      Y/Y %     2009E        Y/Y %
   Product Sales                   139.72               118.85              258.57     139.90              398.47       141.11              539.58
    Svc Rvns                        27.68                25.35               53.04      24.86               77.89       26.00                103.89
 Revenue                           167.40      -26%     144.20     -40%     311.60     164.76       -35%   476.36       167.11      -7%      643.47        -29%

 SEGMENTS                           1Q09        Y/Y      2Q09       Y/Y      1H09       3Q09         Y/Y    9mos09     4Q09E        Y/Y     2009E           Y/Y
 Optical Service Delivery           129.8      -32%      106.0     -48%      235.8        117.1     -41%      352.9     118.7      -13%      471.6         -36%
 Carrier Service Delivery              9.8     -11%       13.0       0%        22.8       23.0       -3%       45.8      22.4       87%       68.2          14%
 Global Network Services              27.8       9%       25.0      -4%        52.8       25.0      -15%       77.8      26.0      -13%      103.8          -7%
   Total                             167.4     -26%      144.0     -41%       311.4      165.1      -35%      476.5      167.1      -7%      643.6         -29%

 Sqntl & Pcntg Rvnu              % Rvnu        Sqntl   % Rvnu      Sqntl   % Rvnu     % Rvnu        Sqntl   % Rvnu    % Rvnu       Sqntl   % Rvnu
 Optical Service Delivery           78%         -5%       74%      -18%       76%        71%         10%       74%       71%         1%       73%
 Carrier Service Delivery            6%        -18%        9%       33%        7%        14%         77%       10%       13%        -3%       11%
 Global Network Services            17%         -7%       17%      -10%       17%        15%          0%       16%       16%         4%       16%
  Total Rvnu                                    -6%                -14%                              15%                             1%
                                                                                                                                                      (continued)
Figure 5.1 (continued)

 WORKBENCH
 Optical Service Delivery         1Q09             2Q09              1H09      3Q09              9mos09    4Q09E             2009E
 CoreDirector                      45.0   -36%       43.0   -48%       88.0      38.0    -52%      126.0     39.5     4%      165.5   -39%
 CoreStream                        26.0   -60%       23.0   -58%       49.0       27.0   -51%       76.0     28.4     1%      104.4   -49%
 Multisrvc Acss T&S (InPho)         1.0   -75%        1.0   -66%        2.0        1.7   -55%        3.7      1.8   -12%        5.5   -57%
 Storage Extension (Akara)          2.0   -50%        2.1   -49%        4.1        2.1   -61%        6.2      2.2   -27%        8.4   -49%
     Total Core                    74.0   -48%       69.1   -52%      143.1      68.8    -52%      211.9     71.8     1%      283.7   -44%
 Metro Optical                      4.0   -67%        3.0   -73%        7.0        3.2   -73%       10.2      3.2   -35%       13.4   -66%
 Cn 4200 FlexSelect                50.0    61%       32.0   -29%       82.0      43.0      4%      125.0     41.5   -30%      166.5    -6%
  Total Metro & Enterprise         54.0    26%       35.0   -38%       89.0      46.2    -13%      135.2     44.7   -30%      179.9   -17%
       Total Transport & Swtch    128.0   -31%      104.1   -48%     232.1      115.0    -42%      347.1    116.5   -14%      463.6   -36%
 Data Networking                   2.00    0.27      2.10    0.27       4.1       2.12   -29%        6.2     2.14     7%        8.4   -30%
 TOTAL Cnvrgd Ethrnt Infra        130.0            106.2             236.2       117.1             353.3    118.7             472.0   -35%

                                 % Rvnu   Sqntl   % Rvnu    Sqntl   % Rvnu    % Rvnu     Sqntl   % Rvnu    % Rvnu   Sqntl   % Rvnu
 CoreDirector                       35%    18%       40%     -4%       37%       32%     -12%       36%       33%     4%       35%
 CoreStream                         20%    -7%       22%    -12%       21%       23%      17%       22%       24%     5%       22%
 Multisrvc Acss T&S (InPho)
 Storage Extension (Akara)
     Total Core                    57%      4%      65%      -7%      61%       59%      -0%        60%      61%      4%      60%
 Metro Optical                      3%    -20%       3%     -25%       3%        3%       5%         3%       3%      3%       3%
 Cn 4200 FlexSelect                38%    -15%      30%     -36%      35%       37%      34%        35%      35%     -4%      35%
  Total Metro & Enterprise         42%    -16%      33%     -35%      38%       39%      32%        38%      38%     -3%      38%
       Total Transport & Swtch     98%     -5%      98%     -19%      98%       98%      10%        98%      98%      1%      98%
 Data Networking                    2%      0%       2%       5%       2%        2%       1%         2%       2%      1%       2%
 TOTAL Cnvrgd Ethrnt Infra                 -5%              -18%                         10%                          1%
                                                Phase 4: The Workbench, Part 1 •   91



Techniques for the Workbench
Units Shipped
The goal of the workbench is to provide more granularity on segment revenue
and margin. Our goal as always is the process, not explication of the theory. So
let’s move on to some examples. In doing so we’ll incorporate all the above-cited
information sources and try to apply them in creative ways in pursuit of greater
accuracy. Here we introduce an important concept: the nature of the market and
the products and solutions offered will partly drive the final revenue tally, par-
ticularly as it relates to the mix of product and services revenue.
       During company presentations, company officers sometimes provide oral,
nonpublished details on segment. Let’s consider a large and diverse (and hypo-
thetical) technology company that participates in multiple market niches, some
serving the consumer and some serving the commercial and industrial markets;
the company is also active in government markets. This particular company on
a recent (hypothetical) conference call confided that it sold 10.1 million home
networking devices in a quarter. The company published the segment revenue.
How do we use this information to better model sector revenue going forward?
       First we’ll want to reference historical conference call transcripts in which
the company provides this information each quarter (again, without publishing
it anywhere else). While there are multiple pay transcript services, and tran-
scripts are included with a Bloomberg subscription, there are also research ven-
dors that publish transcripts soon after the conference call. Some companies
provide transcripts as well. More and more companies are including their pre-
pared remarks as a handout with the results report and presentation.
       At a first cut, the number of units shipped divided by segment revenue
would suggest a wholesale unit average selling price (ASP). But hold on; there will
be a service component. Understanding the product, we know it is a pack-and-
ship product deployed by cable companies and telephone companies. So services
will be small but not immaterial; we can assume 3% to 5%. In determining ASP,
we extract some consistent proportion of revenue from segment sales.
       To model the next quarter, we would use a price based on the modeled ASP,
unit sales, and the service component. But what price? Disaggregate four to eight
past quarters for the segment along these lines. Check the price trend; cross-
check for correlations in inflation gauges such CPI, PPI, and the GDP deflator.
       And how many units? Use the past-quarters’ historical disaggregation to
determine unit trends; look for seasonal patterns. But also use publicly available
data within the GDP accounts and other published federal data on indicators
such as housing creation, incomes, and anything else that might correlate with
92   •   Income Statement Presentation



growth in home networking. Cull the Internet for growth patterns in broadband
access. And assess revenue trends at near rivals.
      Polycom is small enough that it still provides data on unit volumes (some-
thing Juniper used to do but then outgrew). Figure 5.2 shows how we arrived at
our modeled Video Communications revenue estimates throughout 2008. Note
that if our subsegment estimates for a just-reported quarter proved wrong, we
would square them up with the published segment data so that our forward mod-
eling would be based on the most accurate projections.
      Mainly, build the segment revenue model and move on. Don’t agonize over
a fine-tune. Instead, wait for actual results and assess your veracity; how wide was
your miss? Use the miss in subsequent quarters to calibrate and refine future
projections. Here’s a somewhat surprising truth: if your underlying subsegment
assumptions are flawed but you are forecasting accurate segment-level revenues,
keep doing what you’re doing.


Book to Bill
Another hypothetical company forgoes the unit information but does provide
fairly detailed segment book-to-bill information. This might be particularly
applicable for a company that produces lots of relatively inexpensive goods that
do not require financing. Such companies operate a “book-and-ship” business.
Companies producing more complex and customized products likely have some
variant on a “book-and-build” model involving much longer lead times.
      For our book-and-ship company, again we use information from past quar-
terly presentation, in this case to reconstruct an important relationship: prior
quarter book-to-bill ratio in relationship to sequential revenue growth. For our
hypothetical company, we see that in a normal or nonseasonally influenced quar-
ter, book to bill has averaged 1.04 (life is good!), whereas sequential sales growth
has averaged about 0.75% (life is not bad). While those numbers may seem widely
disparate, they actually suggest that the company is converting 96.9% of book-
ings into billings. So to model revenue growth for this particular segment, we
would multiply recent quarter book to bill—let’s say 1.06 (life is really good)
times our multiplier of 96.9%; this figure suggests that sequential sales growth
will be about 2.7%.


Regional Color
Motorola has changed the information it provides regarding mobile devices a few
times in the past decade. At one time, we modeled Motorola’s handset division
Figure 5.2

Polycom provides unit detail on its enterprise video systems and its desktop software-based solutions; using public data, management information, and simple
arithmetic, you can back out the per-unit pricing.

Polycom
Income Statement               1Q08    Yr/Yr %      2Q08    Yr/Yr %      1H08       3Q08    Yr/Yr % 9mos08          4Q08    Yr/Yr %      2008     Yr/Yr %
 Product Revenue                222.45     31%       233.85      16%       456.3     235.40      14%  691.69         222.07      -3%       913.76     13%
 Service Revenue                 36.47     57%        37.74      18%        74.2      40.38     22%   114.58          40.98     17%        155.56     26%
 Net Sales                      258.92     34%       271.58      16%      530.50     275.78     15%    806.3         263.04      -0%      1,069.3     15%

Segments                       1Q08                 2Q08                 1H08       3Q08                9mos08      4Q08                 2008
Product Revnues
 Video Communications            130.3       21%       141.2       25%      271.5     144.2       27%       415.7     141.7       10%       557.4         20%
 Network Systems                  29.2        2%        28.0      -10%       57.2      33.9        4%        91.1      34.2       -4%       125.3         -2%
     Video Solutions             159.5       17%       169.2       17%      328.7     178.1       22%       506.8     175.9        7%       682.7         15%
 Voice Communnications            99.4       77%       102.4       14%      201.8      97.7        4%       299.5      87.1      -11%       386.6         14%
    Total Revenues               258.9       34%       271.6       16%      530.5     275.8       15%       806.3     263.0       -0%     1,069.3         15%

                                                                                                                                                    (continued)
Figure 5.2 (continued)


 WORKBENCH                   1Q08               2Q08              1H08      3Q08               9mos08     4Q08              2008
 Video Comm
 Group Video Units
 (HDX, VSX, Vseries, iPwr_     19,618      9%    20,845      6%    40,463     21,126      8%     61,589    19,831     -6%    81,420    4%
 Rvnue per Unit (000)              60     18%        61     21%                   61     22%                   63     21%
   Grp Vdeo Rvnu              117,218     29%   126,728     29%   243,946   128,437      32%    372,383   124,784     15%   497,166   26%
                                        Q/Q %             Q/Q %                        Q/Q %                        Q/Q %
 Group Video Units                        -7%                6%                           1%                          -6%
 Grp Vdeo Rvnu                             8%                8%                           1%                          -3%

 Desktop Video
 (ViaVideo, PBX Software)     10,703      -5%    11,773     -7%    22,476    12,332      29%     34,808    12,764     17%    47,572    7%
 Rvnue per Unit (000)              1     -88%         1    -89%                   1     -89%                    1    -89%
   Dsktp Vdeo Rvnu               696     -89%       706    -90%                 744     -86%       744        781    -88%     1,525   -87%
                                        Q/Q %             Q/Q %                        Q/Q %                        Q/Q %
 Desktop Video Units                      -2%               10%                           5%                           3%
 Dsktop Vdeo Rvnu                        -89%                2%                           5%                           5%

 Video Prdct Rvnu               117.9    21%      127.4    20%       245      129.2     26%        375      125.6     9%       500    19%
 Video Srvcs Rvnu                13.0    21%       13.4    15%                 15.5     37%                  15.1     9%
    Video Rvnu                 130.9     21%      140.8    20%       272      144.7     27%         416     140.6     9%       557    19%
                                               Phase 4: The Workbench, Part 1 •   95



based on our assumptions about units shipped per technology along with ASP
assumptions per technology. At the time we went to press, Motorola was provid-
ing gross handset unit data (total units shipped worldwide). While it was not
breaking out units by technology (e.g., CDMA, GSM, iDEN), it was providing
unit data on a regional basis. This information was typically provided in public
forums such as the post–results release conference call and in brokerage presenta-
tions, but it did not necessarily make it into SEC documents.
      The regional units data provided by Motorola made it possible to model
total units per quarter, based on prevailing regional trends. We could then arrive
at forecast mobile device segment revenues by using our modeled ASP. In Figure
5.3, we show our actual as well as modeled regional unit assumptions. Based on
our summed regional unit assumptions plus overall unit ASP assumptions, we
can model Mobile Devices segment revenue.


Trade Journal and Industrywide Data
Most industries and many niches are served by one or more trade journals. Trade
journals provide a variety of roles; they contain articles on industry participants,
mainly fawning but sometimes reminiscent of harder-hitting journalism. They
are advertiser supported, and so provide a showcase for new products; they can
also be a forum for experienced industry voices championing or decrying new
developments. Many trade journals additionally provide data—systematic, anec-
dotal, or both—about industrywide products and services, on a unit basis and
sometimes on a revenue basis.
      To make use of such data, you must have an informed opinion about a
company’s individual place in the industries and niches in which it participates.
Let’s assume that you’re modeling a producer of polystyrene, bulk shipments in
a given quarter, and pricing assumptions on the merchant and spot markets. Say
the covered company provides some guidance on revenue from polystyrene in a
given quarter. You can simply model that revenue to grow in line with the com-
pany’s overall forecast growth, adjusted for a seasonal component.
      Or you can actually try to gauge revenue by disaggregating price per unit
from number of units produced. If you are using industry data supplied by a
trade association, you need to have a sense of the company’s competitive place in
the industry. Let’s say that by a combination of sweet-talking IR and utilizing
other industry sources, you determine that the company has a 15% share of the
polystyrene market.
      Let’s begin by using past quarter revenues for polystyrene and the pricing
prevailing in past periods to determine unit volumes. We can then make pricing
Figure 5.3

Motorola discloses regional breakouts of its handset unit shipments, a level of detail that now shapes our Mobile Devices revenue assumptions. In past years, we
had used company-supplied color on technology type (e.g., GSM, iDEN, CDMA 2000) to arrive at our quarterly handset unit forecasts. Our handset unit projections
made at mid-year for the second half of 2009 subsequently proved too aggressive, as the company shifted from a unit-volume focus to a smartphone-based, profit-
per-handset focus.

 Motorola Inc.
 Income Statement           1Q09        Y/Y %      2Q09       Y/Y %      1H09       3Q09E      Y/Y %   9mos09        4Q09E      Y/Y %      2009E         Y/Y %
 Revenue                      5,371       -27.9%     5,497     -32.0%     10,868      5,604     -25.1%   16,472        6,203      -13.1%     22,675       -24.8%

 Segment Sales                1Q09                   2Q09                   1H09     3Q09E                9mos09       4Q09E                  2009E
 Mobile Devices                1,801     -45.4%      1,829     -45.1%       3,630     1,931      -38.0%     5,561       2,215      -5.7%       7,777      -35.7%
 Home & Networks               1,991     -16.4%      2,001     -26.9%       3,992     2,051      -13.4%     6,043       2,122     -18.3%       8,164       -19.1%
 Enterprise Mobility          1,599      -11.5%      1,685      -17.5%      3,284     1,636      -19.4%     4,920       1,881     -15.1%       6,801      -16.0%
 Other                             -       0.0%           -       0.0%          -          -       0.0%         -           -       0.0%           -         0.0%
 Eliminations                   (20)     -50.0%        (18)    -43.8%        (38)       (14)     -59.9%      (52)        (16)     -37.8%        (68)      -48.8%
 TOTAL                        5,371      -27.9%      5,497     -32.0%      10,868     5,604      -25.1%    16,472       6,203     -13.1%      22,675      -24.8%

    Handset Units             14,700     -46.4%     14,800     -47.4%     29,500      15,551     -38.8%     45,051     17,747      -7.6%      62,798       -37.3%
      Revnues Per Unit         117.78      -1.4%    123.58       4.7%                 124.20       1.0%                124.82       1.7%
 TOTAL                          1,731    -47.2%      1,829     -44.9%      3,560       1,931     -38.2%      5,492      2,215      -6.0%       7,707      -36.2%
  Sqntl MOT unit price

                                                                                                                                                       (continued)
     Figure 5.3 (continued)

      Motorola Inc.
      Income Statement        1Q09      Y/Y %      2Q09      Y/Y %      1H09       3Q09E      Y/Y %    9mos09     4Q09E      Y/Y %    2009E      Y/Y %
      North America             8,379    -40.1%      8,673     -41.9%    17,052       9,193              26,245     10,480              36,726
      South America             3,087     -51.1%     3,234    -50.0%      6,321       3,493               9,814      4,086              13,900
      EMEA                        735    -66.5%        441    -73.9%       1,176        454               1,630        504               2,134
      Asiac-Pacific              2,499    -49.4%      2,352    -53.6%      4,851       2,411               7,262      2,676               9,938
          Units Total          14,700    -46.4%     14,700    -47.8%     29,400      15,551              44,951     17,747              62,698

                                            Q/Q    % Rvnu        Q/Q    % Rvnu      % Rvnu       Q/Q    % Rvnu     % Rvnu       Q/Q    % Rvnu
        NA % sales               57%       -23%      59%          4%      58%         59%         6%      58%        59%        14%      59%
        SA % sales               21%       -23%      22%          5%      22%         22%         8%      22%        23%        17%      22%
         EMEA % sales             5%       -23%       3%        -40%       4%          3%         3%       4%         3%        11%       3%
        A-Pac % sales            17%       -23%      16%         -6%      17%         16%         2%      16%        15%        11%      16%
          Units Total                      -23%                   0%                              6%                            14%

      Global Handset Share      5.8%                 5.8%
      Implied Global Units    253,448              255,172
       Sequential Global
       Sequential MOT units
97
98   •   Income Statement Presentation



assumptions based on developing supply and demand trends, which will affect
raw materials costs. (Petroleum-sensitive chemicals will move around a great deal
in cost and price based on changes in petroleum prices.) Once we’ve used back
periods to accurately calibrate pricing and unit volumes, we can use our assump-
tions about go-forward pricing and unit volume outlook to model a revenue fig-
ure for the company’s polystyrene business.


Adjustments to Cost Items
When we are satisfied that our top line for the period reflects our best estimate,
we’ll return to the individual quarters and perform a final tweak of cost items
(this applies to standard P&Ls, not to percentage-of-difference P&Ls). The gran-
ular revenue data provides good insight into cost items that we had previously
treated as simple percentage-of-revenue calculations.
      The first order of business is incorporating any company information, and
this data can be extensive. Companies provide varying levels of disclosure on
near periods and full fiscal years, and they can provide particular guidance on
gross margin and operating cost.
      Nonetheless, the modeler should also work with what the model provides.
In our segment revenue presentation, we also weighted each segment in the
revenue total. This information can shape the gross margin model. Let’s take a
company for which a consumer-sensitive segment represents 60% of revenue;
a segment with a mixed consumer and industrial customer base represents
another 15%; and a segment with a pure industrial customer base represents
the final 25% of sales. Based on our knowledge of the industry and the com-
pany, we expect purely consumer-driven businesses to have 30% gross margin,
while industrial units generate 45% gross margin. Using a simple formula—
(.60 .30)1(.15 .375)1(.25 .45)—we arrive at a weighted gross margin of 35%.
Calibrating our estimate against the actual gross margin, once reported, will get
us closer for the next quarter.
      Another necessary adjustment is to replicate the seasonal flavor of the cost
structure. There are various methods for injecting seasonal adjustments into
costs. These range from sophisticated time series methods such as ARIMA-12
developed by statisticians for the U.S. government, to simple visual confirmation
of past patterns. In an ideal world, we might apply sophisticated formulas to
modeling every cost item. But remember that your investment clients are not
paying you to exactly model R&D; they’re more concerned with your getting the
bottom line right, and most concerned with your getting the valuation right.
                                                 Phase 4: The Workbench, Part 1 •   99



      Companies trade on earnings, or rather the pattern of past, present, and
future earnings. Given earnings’ preeminence, and thus the need to model them
with some precision, our above discussion of seasonal adjustment of cost items may
seem cursory. In my experience, seasonal adjustment of costs can be illusory. As we
develop below, much of the seasonal component of cost adjustment is actually more
tied to revenue than the period. We can’t assume every Christmas will be a home
run for retailers; and we can’t assume that an identically successful year will always
prompt the same bonuses in the fourth quarter for the sales team.


Cost Items: R&D vs. SG&A
Costs at publicly traded firms can be fairly static on a quarter-over-quarter basis,
once a seasonal tweak is assumed. Radical changes in costs on a sequential basis
are rare, unless a company is in the midst of a significant realignment. Every
employee at every company, every process, every site—all are entrenched and
have a shouting interest in staying put. When companies begin cutting costs in
trying times, or conversely adding to operating infrastructure in prosperous
times, the changes begin gradually and often move at a predictable pace, despite
grandiloquent claims to the contrary.
      Given the immense sweep of our project and just a few hundred pages in
which to get it done, we acknowledge our occasional penchant for short-cuts.
Here’s one for the principal cost items: it is typically easier for a company to vary
the SG&A line than the R&D line, on a quarter-over-quarter basis. On a year-
over-year basis, however, we may see more variability in R&D.
      R&D spending reflects costs of developing new ideas, creating and proto-
typing new products and solutions, producing new and improved iterations of
existing products, and determining how to achieve these objectives in a profitable
way. Every project once undertaken has an embedded cost that grows as the proj-
ect moves along, making it that much harder to abandon as it moves (however
unsteadily) toward fruition. Each project brought to fruition represents an incre-
ment to revenue.
      Accordingly, even companies in cost-cutting mode find it hard to cancel
projects in which they’ve invested and that beckon with a payout just a few (weeks,
months, quarters) away. Immediately after a company announces a cost-reduc-
tion initiative, the R&D line may not budge much. But planned new products
may be shelved; and as existing projects mature, the engineers associated with
that project may be laid off. Accordingly, over an extended period R&D will
begin to work down.
100   •   Income Statement Presentation



      SG&A, on the other hand, feels the axe quicker when cost actions are
demanded. When an available market shrinks, fewer sales personnel are required.
Compensation per sales person declines, and performance bonuses decline even
more dramatically. There are long-tailed effects in the SG&A line as well, includ-
ing costs for IT systems. But there is simply more immediately variable cost in the
SG&A line. As conditions and demand improve, the SG&A line tends to inflate
more quickly than R&D for all those same reasons.


Operating Leverage
The factor that in my experience has the deepest effect on operating costs is not
seasonality nor the suppleness of R&D or SG&A spending, but operating lever-
age. This is defined as the amount of change in operating income resulting from
a change in revenue.
      Our mission, to reiterate, is not to diagram theory, not to explicate, defend,
or debunk it, but to grab and use it. The cut-down-to-size theory on operating
leverage is that it is higher for companies with high fixed costs than for compa-
nies with high variable costs. This suggests that high-variable-cost companies are
more nimble in a downturn. The corollary is that high-fixed-cost companies,
after suffering in a downturn, should be able to utilize their fixed costs more
effectively in an economic up cycle, dropping a higher proportion of revenue to
the bottom line than high-variable-cost companies can do.
      What this theory does not account for is that companies do not operate in
a static environment while watching the economic cycle turn up or down. In an
example of cognitive dissonance, we acknowledge Schumpeter’s creative destruc-
tion while averting our eyes to what it does to the fixed-cost company. In his work
Capitalism, Socialism and Democracy, Joseph Schumpeter built on the work of
Werner Sombart to describe the disruptive innovation necessary to sustain and
nourish capitalism. The cost of new ideas and innovation is the demise over time
of entrenched interests and their displacement by new waves of entrepreneurs.
      Creative destruction is most intense in periods of upheaval. Downswings in
the economic cycle—which tend to be more abrupt than the painful and pro-
longed climb back that is the up cycle—are particularly good at scattering verities
in the wind. The fixed assets of high-fixed-cost companies come out of every
down cycle more timeworn, more shopworn, but mainly more competitively dis-
placed than they were going in. We model operating leverage in our model. But
we give due respect to its predictive deficiencies as well.
                                              Phase 4: The Workbench, Part 1 •   101



      In time, you will come to distinguish the variable cost from the fixed cost
in any cost structure and make the necessary adjustments. Our treatment of cost
modeling may seem cursory compared with the time taken with revenue. In our
experience, the top line ultimately drives the bottom line more than the operat-
ing structure. Paradoxically, costs are both more predictable than sales and yet
more specific to the individual company’s operating structure and philosophy
      While the top line is reflective of company-specific practices and processes,
revenue is equally (or perhaps more) influenced by the economic cycle, by new
and existing competitors, and by a host of measures beyond the company’s ken.
While this adds complexity to top-line modeling, it also enables the modeler to
borrow from the broader trends in setting his or her top-line estimates.
      Operating costs, by contrast, are firmly in the company’s control; that
means they tend to be particular to that company. You will learn over time that
granular modeling of individual cost items for a specific company requires famil-
iarity with that company’s long-run operating strategy as well as here-and-now
tactics, which can range from compensation methodology, to litigation appetite,
to adherence to six sigma, to unusually high seasonality . . . and so on. Because
the approach to operating costs is firmly in the company’s hands, the modeler
over time needs to build an operating cost model that is attuned to the individual
company.
      What happens when your finely tuned financial model is upended by, say,
a stock split? How about if your covered company decides to move a substantial
business into a joint venture (JV), and that revenue is (or is not) consolidated on
the other JV partner’s top line? What if you need to model an overseas’ competi-
tor that reports in a foreign currency? Those are some of the topics in our next
chapter.
This page intentionally left blank
                                                    Chapter 6
           PHASE 4:
           THE WORKBENCH,
           PART 2




Special Considerations: Ericsson
Equity Income, Foreign Companies, Stock Splits, and More
We’ve covered most of the techniques required to model the income statement to
a fairly granular level. Inevitably, special situations and exceptions arise that slow
the modeling process. We’ve tried to anticipate some of the situations in this sec-
tion, which we’ve built around a single company. The domestic analyst, in seek-
ing the most comprehensive picture of an industry, will frequently determine that
one or more foreign companies belong in coverage.
       We’ve only briefly touched on equity income and minority interest. More-
over, just when you think your model is good to go, companies have a tendency
to suddenly change their income statement presentation; leading examples
including stock splits and reverse splits along with new segment and revenue
presentations.
       In the following pages, we’ll examine and adjust the model for LM Ericsson
Telefon (stock symbol: ERIC), a company that hits all those notes. Swedish-based
Ericsson reports in local currency but trades ADRs on a major U.S. exchange; has
long had one significant unconsolidated joint venture and now has another;
changed its segment presentation; and engaged in a reverse stock split. We’ll walk

                                                                                  103
104   •   Income Statement Presentation



through the “old” Ericsson P&L presentation on the way to getting it into shape
as the “new” Ericsson.


Modeling Foreign Companies
The topic of modeling foreign companies could certainly fill a book of its own.
Nevertheless, a few basics apply. First, we always model the financial statements—
the income statement, balance sheet, and the cash flow statement—in the native
currency. This is actually a great deal simpler than it was in the days of French
Francs, Spanish pesos, and Italian lira. Now, most of the major European nations,
and thus many of the major companies of interest, report in euros and thus are
best modeled in euros. Sweden has maintained its own currency, and we’ll model
Ericsson in kroner.
      Second, be aware of the differences between U.S. GAAP and the prevailing
foreign accounting standards. Again, the adoption of International Financial
Reporting Standards—IFRS—is making things more uniform and thus making
comparisons across boarders easier and more fruitful.
      Finally, we’ll make the presumption that any foreign-domiciled company
you are modeling has an American Depository Receipt (ADR) that trades actively
on a major U.S. exchange. Most foreign companies of interest to the U.S.-based
analysts trade ADRs on the NYSE or Nasdaq Exchange (a few exceptions, such as
Samsung and BMW, come to mind). The same holds true for depository receipt
equivalents of major U.S. companies that trade on foreign exchanges. If we were
modeling overseas, we would model the financial statements of these U.S. com-
panies in U.S. dollars.
      The principal challenge with foreign modeling comes with valuation, and
we’ll deal with that in our valuation discussion. But in advance of these later
considerations, we’ll make a few adjustments in the foreign-company model that
will facilitate our future valuation analysis in the local currency rather than in
the native currency.
      LM Ericsson Telefon is the leading worldwide provider of communications
infrastructure gear for wireless networks. With an approximate 40% market share,
Ericsson should be followed and at least needs to be acknowledged by any analyst
covering the communications equipment industry. Ericsson is based in Sweden, the
largest industrialized continental European nation not to adopt the euro. Ericsson
reports its results in Swedish kroner; while Ericsson issues translated results and
even conducts its conference calls in English, it does not report in dollars.
                                              Phase 4: The Workbench, Part 2 •   105



The Translation Decision
To model the Ericsson P&L, which is based on IFRS, we’re going to proceed as we
would with any U.S.-domiciled company: use the template of past quarters and
the company’s own presentation style to build the original model and then amend
for a concurrent pro forma presentation. Ericsson provides both segment revenue
and segment operating income, enabling a percentage-of-difference approach. So
far we’ve resisted the temptation to model up from segment operating income,
given that there are so many other things happening in the model. We model
COGS and the operating cost line items in Ericsson’s krona-based results on a
fairly straightforward percentage-of-revenue basis, and link line items sensitive
to the financial structure (i.e., interest income and cost) to the balance sheet.
       The adjustments to the Ericsson P&L are fairly simple but can have signifi-
cant implications if they are mismanaged. The U.S. investor in an overseas com-
pany usually wants dollar translations on two key items: revenues and earnings.
Above the krona-based revenue line, we insert a line and include a translation of
all actual and forecast revenue into dollars. Down below the pro forma diluted
EPS line, we insert a line for dollar-based earnings for the ADR. Finally, remem-
ber that we’ve allowed a line for dividends, even though these are not part of the
formal P&L presentation. Like most European companies, Ericsson pays its divi-
dend semiannually rather than quarterly. We include a dollar translation below
the dividend line.
       It sounds straightforward, but what translation rate do we use? If you’ve
built our recommended model that includes five historical years, two of which
are cut into quarters, you should use company resources or other sources to
determine the historical native currency-to-dollar relationship, expressed in dol-
lars. Foreign companies do not file 10-Qs and 10-Ks but typically file a form 20-F
with the SEC; this is equivalent to a 10-K or annual report. Within the 20-F, the
company will often publish the historical relationship of dollar to native cur-
rency; use those published figures. Lacking that, the Internet can furnish histori-
cal exchange rates.
       Let’s assume you’ve been covering Ericsson for some time now, and the
third quarter of 2009 is winding down. How do we model our currency transla-
tions for the various periods? For the historical first and second quarters (already
reported), we used the historical krona-dollar relationship for each approximately
90-day period. For the third quarter, we use the average rate for the current quar-
ter. And for all forward periods, we use the exchange rate prevailing right this
minute.
106   •   Income Statement Presentation



     While there’s no arguing with the translations for past periods, in the “many
ways to skin a cat” department, analysts adopt various techniques to model future
translations. Some will incorporate formal top-down guidance from in-house or
consensus economists on future exchange relationships. If your firm has the
input of an economist or strategist who has specific forecasts for dollar–native
currency forward exchange rates, you can use those for forward periods. Unless
you can draw on the wisdom of an authority who has been consistently correct
on exchange rate movements (fairly rare), we typically model against the current
exchange rate; at least you know it’s correct today.


Stock Splits and Reverse Splits: The Concept
What do you do if a foreign company’s ADR is not equivalent on a one-to-one
basis with the company’s stock issued in the native country? We’ll also take up
this topic in our valuation discussion. For now, we need to keep the currency
translations correct on a quarterly basis. In the valuation part, we’ll perform the
necessary translation from ADR count to share count.
      For most of the time it has traded ADRs on the Nasdaq Exchange, Erics-
son was one such company. Prior to mid-2008, each ERIC ADR traded on the
Nasdaq was equivalent to 10 ERIC shares traded on the Swedish Exchange. For
that period, the actual Ericsson share base was around 15 billion; calculations
of U.S. dollar–denominated results were converted from kroner and based on
a hypothetical share count of approximately 1.5 billion. On June 2, 2008, the
ERIC shares of LM Ericsson were reverse split at a 1/5 ratio on the Swedish
Exchange. That brought the share base down from about 16 billion at the time
to about 3.2 billion. Commencing on June 10, 2008, the U.S. ADRs of ERIC
ceased trading at a 10/1 ratio to the parent company stock and instead moved
to trading at a 1/1 ratio.
      For U.S. investors, the net effect of these two events was a doubling in the
ADR base in any existing position and a halving in the ADR price. For the Swed-
ish investor, the net effect of these two events was that share positions were pared
to one-fifth their original sizes, and the share price rose fivefold. For the analyst,
the net effect was a familiar progression: a few choice epithets and a sigh of res-
ignation, followed by work to amend the model.
      On the upside, Ericsson’s decision to establish a one-to-one relationship of
ADR to native country stock eliminated a step in our dollar-based valuation
process. It’s fair to say that in squaring its ADR count to its share count, Ericsson
had not the analyst in mind but the investor. In mid-2008, when the transition
occurred, Ericsson was closing out a period of asset acquisitions, including many
                                              Phase 4: The Workbench, Part 2 •   107



U.S.-based assets, as it sought to round itself out from a pure-play wireless infra-
structure company to an integrated network provider. Reacting to its raised pro-
file in the United States, the company wisely aligned its stock with its ADR.
       Stock splits and reverse stock splits follow the fortunes of the market. It’s
fair to say most share-count realignments in the late 1990s and in the 2003–2007
period were stock splits. It is equally fair to say that most realignments in 2000–
2002 and again in 2008 were reverse stock splits, as companies fought to remain
in investors’ view. Many investment portfolios and managers are proscribed from
owning shares priced below a certain threshold; the most common threshold is
$5, though lower (or higher) thresholds are sometimes in effect.
       The official investment theory is that stock splits and reverse splits are
mathematical events with no bearing on a stock’s future trading pattern. If you
can find an investor who actually believes that, you’re Sherlock Holmes. Stock
splits are happy events that tend to draw in momentum investors. Reverse splits
are not death knells, but they do frequently prompt the shorts to get the knives
out. In fairness, I confess to limited familiarity with Swedish investment patterns.
I believe the reverse split in ERIC’s native-country stock was an attempt by the
board to align the ADRs with the stock as a means of encouraging a higher level
of foreign (and specifically U.S.) ownership.


Stock Splits and Reverse Splits: Execution
For the analyst, either kind of split requires some work on the model. This work
is again straightforward, but sensitive; pasting a misplaced “copy” when you
should be pasting a “cut” can cause the model to go haywire. Let’s continue to
work with our ERIC model. And let’s assume we’re maintained the formulas
within the P&L presentation for historical periods. That won’t always be so, inci-
dentally. Try as we may, we can’t always square our pro forma calculations with
a company’s own pro forma numbers; in that case, we may type the company’s
number in the cell. We’ll handle both eventualities.
      When Ericsson decided to reverse-split its stock 5/1, here is how we pro-
ceeded in our native-currency model. We first create a little working room; below
the dividend line and above the cash and debt lines, insert sufficient lines (10
should be sufficient) that are at least equivalent to accommodate diluted and
basic shares, reported basic and diluted EPS, and pro forma basic and diluted
EPS. Because we’re modeling ADRs for a foreign-domiciled company, we need
two more lines, for the dollar translation and the dollar value of native-country
earnings. And we need two further lines for the dividend and the dividend trans-
lated into U.S. dollars.
108   •   Income Statement Presentation



      From the existing P&L presentation, copy the following headers and paste
them into our newly created workspace: basic shares outstanding, diluted shares
outstanding, reported EPS basic, reported EPS diluted, adjusted EPS basic, adjusted
EPS diluted, U.S. dollar EPS, dividends, and U.S. dollar dividends. Let’s work in the
oldest full year, which in this case should be 2007.
      In 1Q07, from the cell showing basic shares outstanding, bring down the
original share count value and divide by 5; do the same for diluted shares. In the
original cell for reported EPS basic, it is not enough to simply copy the cell;
remember, this is a formula cell that divides GAAP net income by basic shares
out. There are a few ways to proceed. You can copy and paste-special the original
values. You can also create a new formula, dividing original GAAP net income
by the new share count, but you’ll eventually need to copy and paste-special this
value before moving it.
      A quick word on what it means to copy and paste-special. When you create
formulas in a cell, and then move those cells further down the page, the cells
referenced in the formula will move as well. For example, if your formula refer-
ences dividing cell A2 by cell A5, and you move your formula 10 rows down the
page, the formula will now reference dividing cell A12 by cell A15. Accordingly,
once we attain a value in a cell that we wish to retain, it is sometimes necessary
to freeze that value and kill the formula in order to retain the value. The proce-
dure is to first “copy” the cell; once we arrive at the appropriate the location,
choose “paste-special”; and from the formula of options provided, choose
“values.”
      So let’s take the original value for reported EPS basic, copy and paste-special
into the new cell, and multiply by 5. One line below, do the same for reported EPS
diluted. Repeat this process for pro forma EPS basic and diluted. For U.S. dollar
EPS, again it makes sense to copy and paste-special the value into the new cell.
Do the same for SEK (Swedish krona) dividend and U.S. dollar dividend. Remem-
ber, this foreign company pays a semiannual dividend and is unlikely to show a
first-quarter dividend. Nonetheless, put these formulas in place for dividend and
dollar dividend as we plan to replicate these cells as part of a “stack” of formulas.
Copy this stack of formulas for all periods and interim periods for the year 2007
and for all historical quarters and years; let’s assume we’re doing this exercise
early in the year 2009 before the 1Q09 report.
      Next, copy and paste-special all these cells to freeze their values. Then, one
year at a time for historical periods, copy and paste the new and correct cells over
the original share counts, reported and pro forma EPS, dollar values and
dividends.
                                               Phase 4: The Workbench, Part 2 •   109



      Figure 6.1 shows Ericsson’s native-currency P&L model for 2008 on a pre-
split basis (taken from a spreadsheet used at the time) and also on a postsplit
basis. While the 2008 second quarter and second half were modeled in the pre-
split presentation, the second-quarter and second-half results are actual in the
postsplit showing. Conditions were deteriorating quickly in the second half of
2008, and we can see that actual results badly missed our forecast if we “eyeball-
adjust” for the 5/1 split.
      What about go-forward periods? Recall that the share count for our first
nonhistorical and modeled quarter, in this case 1Q09, is based on share count for
the preceding (4Q08) period. If the model has been properly built, all future
periods should automatically begin using the new share count. On that basis, all
the per-share calculations going forward will adjust to the new share basis.


Formal Modeling of Equity Income and Minority Interest
Sony Ericsson Mobile Communications (Sony-Ericsson) was established in 2001.
At the time, several European-based handset makers were discovering and con-
fronting the difficulty of competing on a stand-alone basis with leaner, low-cost
Asian operators. With the exception of Nokia, by 2005 all the European handset
divisions—most notably Alcatel, Siemens, and Ericsson—had been blended into
ventures with Asian partners or sold. Ericsson is the only one of those three to
retain any parent-company identity.
       We’ve already touched on equity income and minority interest modeling in
our P&L presentation discussion. For the buy-side analyst or other analyst
charged with modeling a large number of companies, we recommended using
historical relations to model these values. For example, if minority interest typi-
cally equates to 3% to 5% of income, disgorge that amount from each quarterly
model. We have generally found it easier to accurately model minority interest
than to accurately model equity income.
       For the analyst with more support or time, a more accurate input can usu-
ally be derived from modeling the joint venture that contributes to equity income
or subtracts minority interest. This is particularly true if public information is
available to help us. Sony-Ericsson provides detailed information each quarter,
including a brief P&L statement, the number of handsets shipped, and the aver-
age selling price per handset. Early in 2009, Ericsson formally began participation
in ST-Ericsson, a joint venture that competes in communications semiconduc-
tors. Initially, the ST-Ericsson JV provided less financial detail than Sony-Ericsson.
Figure 6.1

In mid-2008 Ericsson reverse-split its shares one for five and split its ADRs two for one at least partly in an effort to align its shares with its ADRs (which formerly
traded ten for one).

Ericsson
Income Statement                 1Q08                   2Q08E                  1H08       3Q08E               9mos08        4Q08E                 2008E Yr/Yr %
Krona/dollar                       5.977                    5.977                 5.977      5.977                5.977         5.977                5.977
 US Dollar sales                    7,391                  7,582                 14,973      7,774               22,747         8,632               31,378    9%
 Net Sales                       44,175.0       4.8%     45,315.8     -4.8%    89,490.8   46,465.2       6.7% 135,955.9      51,590.9      -5.3% 187,546.9   -0%

Pre-Split
Net Income                        2,645.0     -54.8%     4,005.7     -38.5%     6,624.7     5,081.8     28.5%    11,706.5     5,804.6      2.9%      17,511.1     -21%
PF Net Income                     3,445.0      -41.1%    4,005.7     -38.5%     7,424.7     5,081.8     28.5%    12,506.5     5,804.6      2.9%      18,311.1
Basic Shares Outstndng           15,905.0        0.1%   15,944.8       0.3%    15,924.9    15,944.8      0.3%    15,931.5    15,944.8      0.3%     15,934.8        0%
Diluted Shares Outstndng         15,983.0        0.6%   16,023.0       0.8%    16,003.0    16,023.0      0.3%    16,009.6    16,023.0      0.3%     16,013.0        1%
Reprtd EPS Basic                     0.17     -54.8%        0.25     -37.2%        0.42        0.32     28.1%        0.73        0.36      2.6%          1.10     -21%
Reprtd EPS Diluted                   0.17     -55.1%        0.25     -37.5%        0.41        0.32     28.1%        0.73        0.36      2.6%          1.09     -21%
Adjstd EPS Basic                     0.22     -41.2%        0.25     -37.2%        0.47        0.32     28.1%        0.79        0.36      2.6%          1.15     -18%
Adjstd EPS Diluted                   0.22     -41.5%        0.25     -37.5%        0.46        0.32     28.1%        0.78        0.36      2.6%          1.14     -18%
US Dollar EPS                        0.36                   0.42                   0.78        0.53                  1.31        0.61                    1.91      -9%
Dividend                                                                           0.25                              0.25                               0.50

Post-Split
 Basic Shares Outstndng          3,181.00         0%    3,183.00         0%    3,182.00    3,184.00        0%    3,182.67    3,185.00        0%     3,183.25        0%
 Diluted Shares Outstndng        3,196.60         1%    3,226.45         2%    3,211.53    3,184.00       -0%    3,202.35    3,200.00        0%     3,201.76        0%
 Reprtd EPS Basic                    0.83       -55%        0.60       -70%        1.47        0.89      -28%        2.39        1.22      -31%         3.68      -47%
Reprtd EPS Diluted                   0.83       -55%        0.59       -71%        1.45        0.89      -28%        2.37        1.21      -31%         3.66      -47%
Adjstd EPS Basic                     1.08       -41%        0.60       -70%        1.72        0.89      -28%        2.64        1.22      -31%         3.93      -44%
Adjstd EPS Diluted                   1.08       -41%        0.59       -71%        1.70        0.89      -28%        2.62        1.21      -31%         3.91      -44%
US Dollar EPS                        0.18                   0.10                   0.28        0.13                  0.41        0.16                   0.57      -46%
 Dividend                                                                          0.25                              0.25                               0.50
                                              Phase 4: The Workbench, Part 2 •   111



As of this writing, additional detail was being provided, though less than you
would find with an independent public company.


Modeling an Existing Joint Venture
As a continent-straddling joint venture, or JV, Sony-Ericsson reports neither in
kroner nor dollars but in euros. This provides another useful modeling exercise
as we incorporate a third currency along with the kroner and dollars. Before cry-
babying about three currencies, count yourself lucky that ST-Ericsson is not
incorporated in Geneva and reporting in Swiss francs; as it happens, JV partner
ST Microelectronics is based in Switzerland. To complete the trifecta, ST
Microlectronics reports in dollars, because most communications semiconductor
transactions occur in dollars.
      For Sony-Ericsson, for the first time we are going to stray off our P&L pre-
sentation worksheet. As a first step, insert a new worksheet into our workbook,
something that we have not yet done often. We’ll assume that you are sufficiently
familiar with Excel to add a worksheet. Later, we’ll be adding multiple new work-
sheets (and even a linked multicompany workbook); but that’s further along in
the cattle drive.
      On our spanking new worksheet, we will download the historical data avail-
able from Sony-Ericsson. Remember that this joint venture does not have its own
SEC filing obligations. We can find all of the information we need on the Sony-
Ericsson Web site. You can purchase financial statements for Sony-Ericsson from
various fee-based services, which are ready to download into Excel format. Notice
that the Sony-Ericsson P&L does not conclude with per-share data but instead
ends with after-tax earnings for the joint venture. All earnings (losses) are shared
by the partners, and there are no Sony-Ericsson shares trading on any
exchange.
      However you acquire the data, input our usual four to eight historical quar-
ters of essential information. Essentials for this presentation are the P&L and the
usual percentage of revenue items below the income statement. In an abbreviated
workbench, at a minimum we’ll include the number of handset units shipped and
the average selling price. Remember that we are modeling in euros and that all
currency values, including ASPs, will be in euros.
      Those inputs are the absolute minimum you can use to contribute useful
information to your Ericsson consolidated P&L presentation. Because I perform
deeper-dive analysis on Ericsson, my Sony-Ericsson worksheet also incorporates
other information provided either formally or anecdotally by the company,
including revenue by region (reported) and some color on technology type and
112   •   Income Statement Presentation



model types shipped (provided anecdotally in the post–results release conference
call or from other sources).
      To model Sony-Ericsson, we will use percentage-of-revenue inputs for most
line-item cells. To estimate the top line, we multiply handset units times ASPs.
For ASP, we reference nearest-quarter ASP, adjusted up or down. (Like most aver-
age selling prices in the technology sector, handset ASPs are almost invariably
heading down; for a blended mix of handsets, the variable to this near-constant
would be a richer mix of smart phones or other higher-margined units.)
      To derive a forecast for unit shipments, we can build the shipment model
based on our anecdotal information about product types (e.g., Cybershot, Disk-
man, and smart phone). But since this is a merely a supporting worksheet, we can
model unit shipments based on some variation of nearest historical quarter ship-
ments. We also need to be aware of historical patterns; because handset end users
are most often consumers, the fourth quarter will have a bulge not explained by
normal sequential progression.
      As you can see, so far this is nothing more than a cooked-down P&L pre-
sentation. The way it impacts the consolidated income statement is slightly com-
plicated in the case of Sony-Ericsson because of the necessary currency
conversions. Because many joint ventures are domiciled outside the native coun-
try, knowledge of this process will be useful going forward for many other
companies.
      In this model, the adjustments impacting the master consolidated P&L
occur about 100 lines down the Sony-Ericsson worksheet. There are two key
adjustments: (1) the currency conversion and (2) the adjustment to strip out
Sony’s share of the profit (or loss). For the currency conversion, the relationship
we need to model is euros to kroner. It is important that we gauge the accuracy
of past adjustments of Sony-Ericsson data to the consolidated P&L, so I recom-
mend that you obtain the preceding four to eight historical quarters of the rela-
tionship of euros to kroner.
      For the currency adjustment, insert a line within which to convert euros
into kroner at the prevailing exchange rate for each period. As we’ve stated, these
are the historical exchange rate for past periods, average exchange rate for current
period, and current exchange rate for future periods. Let’s begin our work with
a historical period for which we know: (1) the euro/krona exchange rate; and (2)
the actual contribution from the joint venture to the consolidated P&L. For 1Q08,
each euro was equivalent to approximately 9.35 kroner.
      The contribution to the consolidate P&L occurs above the income tax line,
so we will concern ourselves with Sony-Ericsson pretax income. For our histori-
                                                Phase 4: The Workbench, Part 2 •   113



cal period of 1Q08, Sony-Ericsson pretax income was equivalent to 193 million
euros; at the prevailing average exchange rate for that period, Sony-Ericsson pre-
tax income converts to 1.804 billion kroner. Sony-Ericsson is a 50/50 joint ven-
ture; on that basis, half the 1.804 billion kroner, or 902 million kroner, should go
to Sony, while the other half should be contributed to the LM Ericsson consoli-
dated income statement.
      Remember that Ericsson, as a multibillion-dollar organization, may have
more than one JV relationship; some may be so insignificant that they hardly
bear mentioning. The company may have calculated the euro/krona conversion
differently, based on the actual pace of business transactions. Various other
transfer or administrative fees and costs can interfere. For all that, our 902 mil-
lion kroner calculation lines up well with the actual historical contribution of 911
million kroner for 1Q08.
      Figure 6.2 captures a happy year for Ericsson, as the joint venture was attain-
ing record market share in the 9% range, and the parent company was solidly
profitable. For all of 2007, our Sony-Ericsson worksheet indicates a modeled con-
tribution of 7.136 billion kroner; the actual contribution was 7.232 billion. For full
year 2007, our model was 98.7% accurate; for 1Q08, the accuracy was within 1%.
You can’t always count on such good results; indeed, by late 2008, the accuracy had
degenerated. Late 2008 was a period in which formerly highly profitable Sony-Er-
icsson found itself ill prepared for a shift in tastes away from its feature phones and
toward smart phones. The joint venture’s growing losses may have contributed to
the decreasing alignment between modeled contribution from the Sony-Ericsson
sheet and actual contribution on the consolidated income statement.
      What are we to make of this information, and specifically this level of cor-
relation? Generally, I find that modeled Sony-Ericsson pretax income, post–cur-
rency conversion, is approximately 95% accurate. When the joint venture is
running hot (which hasn’t happened in a few quarters), I tend to accord a 5%
premium to the modeled pretax contribution. When the joint venture is forecast
for modest and/or declining profits, I assume that only 95% or even 90% of the
modeled contribution will make it to the consolidated equity income lines. When
I’m modeling a loss (like the experience I had in 2009), I’ll assume the loss will
be somewhat worse by 5% or 10% by the time it reaches the consolidated P&L.
      As for moving this information onto the consolidated P&L, we’ll again
assume some passing familiarity with Excel. On the consolidated P&L, for the
period to be modeled, place an sign in the Equity income cell; then go to the
Sony-Ericsson worksheet and select the appropriate period from the line for 50%
share to Ericsson.
Figure 6.2

Sony-Ericsson’s revenues are consolidated on the Sony P&L, not that of Ericsson. Half the joint venture’s pretax earnings, translated from euros to Swedish kroner,
appear on Ericsson’s P&L as Share in JV Earnings.

Sony-Ericsson
(Euros)
                             1Q07        Y/Y %      2Q07        Y/Y %       1H07         3Q07E        Y/Y %      9mos07       4Q07        Y/Y %      2007         Y/Y %
Sales                         2,925.0       46.8%    3,112.0       37.0%     6,037.0       3,108.0          6.7%   9,145.0     3,771.0       -0.3%   12,916.0        17.9%
COGS                          2,038.7       38.8%    2,192.0       34.9%     4,230.7       2,154.0          8.0%   6,384.7     2,573.0       -4.2%    8,957.7        15.2%
Gross Margin                    886.3       69.1%     920.0        42.2%     1,806.3        954.0           4.0%   2,760.3     1,198.0        9.3%    3,958.3        24.3%
 R&D                            261.0       29.2%     283.0        26.9%        544.0       280.0         24.4%      824.0       349.0       36.3%     1,173.0       29.5%
 SG&A                           284.0       52.7%      321.0       30.5%        605.0       280.0         -2.4%      885.0       375.0        2.2%    1,260.0        16.0%
 Operating Expense              545.0       40.5%     604.0        28.8%      1,149.0       560.0           9.4%   1,709.0       724.0       16.2%    2,433.0        22.1%
 Other Op Lss (Incm)             (5.0)     -28.6%         2.0    -107.7%         (3.0)          1.0    -104.8%        (2.0)     (15.0)       50.0%       (17.0)     -73.4%
Operating Income                  346     142.2%         314       53.9%          660          393         -7.7%     1,053       489.0        1.2%       1,542       22.8%
 Financial Income              (18.0)     100.0%      (18.0)      125.0%       (36.0)         (7.0)      -12.5%     (43.0)      (19.0)        0.0%      (62.0)       40.9%
 Financial Expense                 2.0       0.0%         6.0        0.0%          8.0        16.0    1500.0%          24.0         7.0       0.0%         31.0   3000.0%
Pretax Income                     362     138.4%         326       53.8%          688          384       -11.3%      1,072       501.0       -0.2%       1,573       21.1%
Taxes                             100     194.1%          97       51.6%         197.0         109       -14.2%      306.0       118.0     174.4%        424.0       58.2%
Tax Rate                         28%        23.4%       30%         -1.4%        29%          28%         -3.2%       29%         24%      175.0%         27%        30.6%
 Minority Interest                   9       0.0%          10     100.0%          19.0            8         0.0%       27.0         10      -16.7%         37.0       8.8%
Net Income                        254     133.0%         219       53.1%          472          267       -10.4%        739         373      -16.6%        1,112      11.6%
      Figure 6.2 (continued)


      Sony-Ericsson
      (Euros)
                               1Q07        Y/Y %     2Q07        Y/Y %     1H07        3Q07E        Y/Y %     9mos07       4Q07        Y/Y %      2007        Y/Y %
      Euro-Krona Conversion       9.390                 9.390                               9.114                              9.114
      PreTax Incm Converted     3,249.0              2,957.90                           3,499.58                           4,565.86
      50% Share to ERIC        1,624.49              1,478.94              3,103.43     1,749.79               4,853.22    2,282.93               7,136.15

      Ericsson
      Income Statement            1Q07     Yr/Yr%       2Q07     Yr/Yr%       1H07        3Q07      Yr/Yr%     9mos07         4Q07     Yr/Yr%         2007    Yr/Yr %
      Krona/Dollar                 6.698                6.698                 6.698       6.407                    6.601       6.303                  6.526
       US Dollar sales             6,294                 7,110               13,404        6,797                  20,198       8,640                28,773
       Net Sales                42,156.0      6.5%   47,619.0       7.8%   89,775.0    43,545.0        6.8%   133,320.0    54,460.0       0.5%   187,780.0        5%
       COGS                    24,034.0         8%   27,166.0       6.1%   51,200.0    28,050.0       10.3%    79,250.0    34,809.0      11.1%   114,059.0        9%
       Gross Profit              18,122.0        5%   20,453.0      10.2%   38,575.0    15,495.0        1.0%     54,070.0    19,651.0    -14.1%     73,721.0      -0%
       R&D                       6,453.0       -3%    7,208.0       5.1%   13,661.0     7,229.0        2.2%    20,890.0      7,952.0      9.9%    28,842.0        4%
       SG&A                      5,322.0       11%    5,856.0      11.3%    11,178.0    4,783.0       -9.7%     15,961.0     7,238.0     19.2%     23,199.0       8%
       All Other                 (162.0)       41%    (389.0)     -52.4%     (551.0)    (402.0)      -89.3%      (953.0)     (781.0)     -8.0%    (1,734.0)
       Share in JV Earnings    (1,642.0)     136%    (1,477.0)     48.9%   (3,119.0)   (1,751.0)     -14.0%    (4,870.0)   (2,362.0)      6.9%   (7,232.0)
115
116   •   Income Statement Presentation



Modeling the New JV
In early 2009, Ericsson and ST Microelectronics formed a JV company that con-
tained the communications semiconductor businesses and related assets of the
two parent companies. Like Sony-Ericsson, this business is not consolidated on
the Ericsson top line. Instead, it will contribute equity income (or loss) to Erics-
son. By way of background, ST Microelectronics had only recently concluded the
acquisition of Royal Philips’ semiconductor assets when it agreed to combine its
communications semiconductor business into the ST-Ericsson joint venture. The
combination of these established business units has so far been balanced by the
predictable disruption of combination and integration; the unit also began life in
the worst demand climate for handsets and thus for communications semicon-
ductors in a decade.
      For the modeler, the takeaway is that this unit is in its early days only a
modest contributor or detractor from overall equity income. Its contribution (or
decrement) is currently swamped by the much more volatile contribution from
Sony-Ericsson. But keep in mind that on the Ericsson P&L presentation the two
contributions will not be disaggregated but will be presented as one number.
      In the very early phase, we did not create a separate worksheet in the work-
book to model ST-Ericsson in detail. Initially, with little historical data to build on,
we used a “plug” number to reflect the (loss) contribution from ST-Ericsson and to
adjust the equity income line that was dominated by Sony-Ericsson. As we were
wrapping up preparation of this book, the joint venture gathered some momentum
and built some history; at that point it earned its own detailed worksheet. And the
Ericsson equity income line began to reflect contributions (in each case, losses)
from the two ventures. These were dollar-based losses from ST-Ericsson and euro-
based losses from Sony Ericsson, both translated into Swedish kroner.
      There is another key takeaway for the modeler who began to follow Ericsson
in the period after the Sony-Ericsson inception but before the ST-Ericsson launch.
That is, the revenue from the unit formerly known as Ericsson Mobile Platforms
(EMP) is now contributed to the ST-Ericsson joint venture and accordingly no
longer figures in the Ericsson consolidated top line. The point is that the Ericsson
handset revenue and contribution has long been absent from the consolidated
P&L. Modeling the new joint venture requires more than calculating the effect
on the equity income line, as we routinely do with Sony-Ericsson. To accommo-
date the effects of ST-Ericsson, the modeler must simultaneously calculate its
effect on equity income and disaggregate the effects of the absent business (in this
case, EMP) from the parent company’s consolidated financial statements.
                                              Phase 4: The Workbench, Part 2 •   117



New Segment Modeling Presentation
Ericsson, a large company, is generous in filling all our special-interest needs. So
far it has helped us by trading ADRs while modeling in a foreign currency,
reverse-splitting its stock and splitting its ADRs, and participating in an existing
and now a new JV company. The company has also sold and (mainly) acquired
assets. Most of the asset acquisitions were intended to transform the company
from a pure-play wireless infrastructure equipment company to a provider of
converged next-generation network solutions.
       Whenever this fine-sounding gobbledygook is ringing out, you can bet
there is a cadre of analysts somewhere tearing their hair out. That’s because seg-
ment reorganization can completely shatter a carefully crafted P&L presentation.
Sometimes a change in reporting segments reflects the maturation of a company;
this is particularly true in a dynamic segment such as technology. Almost from
inception, Juniper Networks during its post–results release conference call would
provide very granular details on its operations. Specifically, the company would
disclose the number of router chasses sold in the quarter; it would also disclose
the number of blades, which are the high-value-added products that are slotted
into an empty or partially filled chassis. As Juniper Networks matured from a
$100 million revenue company in 1998 to a $3.5 billion revenue company by
2008, analysts recognized that it was only a matter of time before this informa-
tion would stop being available; that happened sometime in 2007.
       General Electric (GE) has changed its segment presentation repeatedly over
the years. For GE, we have always felt that each new segment presentation pro-
vided a kind of mirror into managements’ strategic thinking. For the GE finan-
cial part of the business, the company used to hold as many as 25 competencies
in five broad areas: equipment leasing, insurance, consumer finance, commercial
finance, and diversified. The company eventually boiled down to a tighter focus
in each of these areas; realignment of GE Capital is an ongoing process and will
likely always remain so. For industrial GE, formerly disparate lines such as light-
ing and appliances were either divested or folded into a broad Consumer category,
itself slated to go away. Jet engines and gas turbines lost their headings and were
grouped into a Power unit; the Infrastructure unit gathered formerly disparate
businesses into a niche of its own.
       Our concern, as always, is not with corporate motive and strategy, but with
how we reflect the changes and—importantly—whether and how we capture and
save the obsolete segment information. We have already discussed the mechanics
of adjusting a model to accommodate a stock split or reverse stock split. Any P&L
118   •   Income Statement Presentation



presentation that draws inputs from below the margin presentation must be
amended carefully to ensure that key inputs do not go awry or reflect out-of-date
modeled segments. For that reason, it is probably a good idea to remove the obso-
lete segment representation from the P&L worksheet as soon as the new segment
data become available.
      At the same time, the old segment data often contain useful information
that might inform or color our analysis going forward, even if that analysis is
based on a new data set. Given that consideration, it makes sense to save a copy
of the old workbook that includes the historical and now obsolete segment
presentation.
      Later in our discussion, in Part 4, we will describe how to create a matrix
workbook that incorporates data from various individual company workbooks
in order to analyze industry participants and data in aggregate. This matrix
workbook will have multiple links to the individual company workbooks.
Accordingly, once the matrix workbook is created, we need to be careful about
simply “saving as” copies of old presentations and filing them away. We need to
be mindful of creating the new segment presentation and segment model on the
individual company workbook that is linked to the matrix workbook, and not
inadvertently stashing the still-linked workbook in a folder labeled “old stuff” or
the like.


Procedure
As noted, beginning in the middle of the decade, Ericsson made a series of asset
acquisitions principally in North America that were intended to transform the
company from a pure-play wireless infrastructure equipment company to a pro-
vider of converged next-generation network solutions. Historically and through
most of 2007, the company reported both mobile networks and wireline network,
which were summed in total network equipment; network rollouts were pre-
sented as a subset of this total. Ericsson also reported professional services and
other services. Ericsson also provided operating income per segment.
      Late in 2007, the company announced that new segment reporting was
imminent. And beginning in 2008, Ericsson started to report network equipment
as a single figure, while continuing to break out network rollouts as a subset of
this total. The company continued to report professional services, but added
managed services as a subset of this total. And the “other” category was replaced
with multimedia. As in the past, Ericsson furnished operating income per seg-
ment for the new reorganization.
                                               Phase 4: The Workbench, Part 2 •   119



      As is often the case, the company used the segment reorganization to shift
assets from one broad heading into another. So, while you might expect services
to be unaffected by its new title, for 2006, recalculated professional services for
2006 was 14% larger than the historical representation of services used prior to
2008. It follows that segment income was also altered by the reallocation of assets
under the new headings.
      With all this in mind, and assuming the links to the matrix workbook are
already established, here’s how we proceed upon learning that Ericsson has
changed its segment reporting. Study the new presentation and determine how
much space you’ll need in your existing Ericsson workbook. In this case, you’ll
need quite a lot, about 50 lines. That’s because we need to accommodate the new
presentation of segment revenue, with annual comparisons side by side; our stan-
dard percentage-of-revenue and sequential change representation, which repli-
cates the segments and takes up equal room immediately beneath; the operating
income presentation; and the segment margin presentation. (I typically dispense
with a percentage-of and sequential change representation for operating
income.)
      Add as much restated historical data as the company provides in the new
representation. Typically, the company will go back a complete year by quarters
as well as any quarters reported for the current year. Use the restated segment
data, as well as any contributing data from your workbench below, to model the
new segment presentation.
      Once you have a complete set of percentage-of and sequential change rep-
resentations for the restated year, copy and paste them below each to-be-modeled
segment presentation. Use the historical segment income data and segment mar-
gin information, along with information about current conditions, to model
income for each segment going forward.
      The next step is delicate and is the place where the model can go awry. At
the very top of the P&L presentation, in the modeled consolidated revenue cells,
you need to substitute your new modeled segment total to replace the old segment
totals (still residing beneath the new segments). Remember that linked cells don’t
show their links visibly. If you fail to carefully reassign the links, you’ll have no
visible indication that your crucial top-line model is drawing from the wrong
data source.
      That’s why we recommend eliminating the old segment presentation once
the recalculated segment data has been inserted and the new segment presenta-
tion is now the linchpin of our top-line model. Keeping in mind the need to
preserve links to the matrix (all-company) workbook, at this stage we recom-
120   •   Income Statement Presentation



mend “Saving As” a copy of the Ericsson workbook with a designation in the title
(such as “old” or “defunct”). Close the workbook. Store the workbook with the
“defunct” designation in a file folder created specifically for stashing old data that
you won’t actively need but may need to source some day.
      Open the original Ericsson workbook that has links to the matrix work-
book. Only at this point will we eliminate the obsolete segment presentation. We
do so for two reasons. One is clarity; we want to maintain only that data on the
model that is actively useful for our modeling purposes. The other is as a fail-safe.
If you eliminate the obsolete presentation and the model goes haywire, it means
you failed to fully sever the inputs to the top line coming from the obsolete data.
If you encounter such an issue, which will sometimes reduce all future quarters
to chaos, don’t panic. Coolly check the top-line inputs to make sure that they are
aligned with the new segment presentation. That will solve your problem.
      We’ve now covered some of the special situations—foreign companies,
stock splits, treatment of JV-related minority interest and equity income—that
commonly confront the analyst. With these specials and exceptions out of the
way, we’re nearly done with our income statement modeling overview. But before
we finish, let’s use historical and modeled data to prepare normalized or “eco-
nomic-cycle-free” earnings. And let’s refine our approach to historical growth to
accommodate a more representative data set.
                                                  Chapter 7
           ORDINARY LEAST
           SQUARES REGRESSIONS
           AND NORMALIZED
           EARNINGS




Static Growth, Dynamic Cycles
If companies behaved themselves—always grew the top line, carefully managed
costs to prudently expand margins, and delivered ever-improving net income—
the valuation process would be infinitely simpler. They don’t, of course; and for
the course of business, that’s a good thing. We need Schumpeter’s creative
destruction to invigorate the economy and ensure a healthy, inclusive, and
expanding capitalism. Innovation is the gardener, constantly turning over the
competitive landscape; last year’s blue ribbon winner is this year’s mulch.
       And then there’s the economic cycle, which moves across our landscape
metaphor like a cyclone; it’s an equal-opportunity town leveler. The business
“cycle” is a misnamed phenomenon, if ever there was one, because the business
cycle is anything but cyclical. Once you are within a business cycle, you can
expect the economy to work through various phases, albeit erratically: growth,
hypergrowth, downturn, trough, stabilization, early recovery, and growth once
more. But the business cycle does not arrive cyclically, like spring following win-
ter; it tends to hit an individual company’s earnings cycle like a car crash. After
the crash, the gawkers will say they “saw it coming.”


                                                                                121
122   •   Income Statement Presentation



      The healthy chaos of competition creates challenges to the growth modeling
and valuation process. To derive valuation assumptions based on forward-look-
ing inputs, we need those inputs to be as reliable as can be; equally important, we
expect them to be as economy sensitive as they can be. Yet analysts and investors
are wedded to a five-year forecast of annualized EPS growth. And they tend to
base forward-growth assumptions on historical growth rates or that five-year
hypothetical rate, regardless of the economic climate.
      Analysts worth their salt adjust forward assumptions based on necessarily
subjective perceptions, such as management performance and stability, dynamic
positioning within dynamic end markets, financial health, quality of customers,
and the like. For all that, and even if we bake in variations based on our percep-
tions of these subjective circumstances, growth assumptions lean heavily on his-
torical growth patterns and performance.


Shortfalls of Unadjusted CAGRs
The problem is that historical growth itself is dodgy. What’s the appropriate his-
torical measurement period? More significantly, how do we measure? To get a
five-year look-back on EPS growth, some analysts will measure five annual rates
of change and then average the five rates of change. That’s a far-from-optimal
method, we believe, in that it equal-weights typically faster-growth earlier peri-
ods with slower-growth later periods, thus distorting the actual trend. More
commonly, analysts will use a multiperiod compound annual growth rate
(CAGR) formula, which incorporates each year’s growth rate into the final tally
and likely gives a more “honest” reading.
      The chief shortcoming of unadjusted CAGR, however, is that it relies on two
single data points: the initial-period data, and the end-period data. What would
happen if company-specific strategy or economic cycle gyrations distorted one
of those inputs?
      Consider a hypothetical company that earns $1.00 per share in 2004 and
grows its earnings 20% annually on average from 2004 through 2007. In 2008, in
response to aggressive competitors, the company meaningfully ramps its R&D
development costs, attacks the market more aggressively with its sales effort, and
slashes prices to win share in fast-closing emerging markets; earnings for the year
total $1.10. At the tail end of 2008, management reassures investors that resump-
tion of its normal go-to-market, pricing, and operating practices will restore
growth in 2009. On the bottom line, you take that to mean 2009 EPS will be
                   Ordinary Least Squares Regressions and Normalized Earnings   •   123



building off a base somewhere between the $1.74 EPS earned for 2007 and the
$1.10 earned for 2008.
      Based on the $1.00 earned in 2004 and the $1.10 earned in 2008, the stan-
dard CAGR calculation shows the company’s five-year annual growth rate as
1.92%. Your model is structured to use the five-year CAGR to project forward.
Based on all other available data and your knowledge of the company and indus-
try, 2% forecast growth meaningfully understates operating prospects. You can’t
simply plug in the prior 20% growth rate and act as though the 37% single-year
drop (from $1.74 in 2007 to $1.10 in 2008) didn’t happen.


Ordinary Least Squares (OLS) Regressions
You can, however, get a much truer assessment of the underlying trend if you
incorporate data from all the individual periods. By its nature, a CAGR calcula-
tion must use beginning and end points. But those two points can be adjusted
based on the data flow across the individual periods. The method we use to
incorporate all of the pertinent annual data points is ordinary least squares
regression (OLS growth, for short). Using OLS growth, CAGR for the hypotheti-
cal company and the period described above is 4.6%.
       That may or may not be your final EPS growth forecast, but it is a more
honest and certainly more comprehensive appraisal of the five years just con-
cluded. As a look-back mechanism, OLS can appropriately size individual com-
pany events that by Wall Street convention were not eliminated from pro forma
calculations. We believe OLS is particularly useful in mature industries where pro
forma calculations are not commonly used to adjust out one-time events. Perhaps
its biggest value, however, lies in its ability to wrest a growth rate when one of the
end points is negative
       It is difficult, if not impossible, to predict when a company will make a sud-
den veer in its operating philosophy or marketing strategy. No less difficult to
predict, but inevitable in outcome, are changes in the economic cycle. A further
merit of the OLS is that it can reach back through all aspects of one or more
economic cycles to give a more comprehensive picture of long-trend earnings.
While a long-term simple CAGR could do that as well, it is again subject to the
vagaries of the end points chosen.
       Well and good for historical data, but what about the economic surge or
decline that is five years, or two months, away? Should we blithely issue EPS
projections for coming years, ignoring the chance that a recession could chop
124   •   Income Statement Presentation



them down like corn rows before an Iowa twister? Why, yes. Remember, the
market trades on collective perceptions, not our superior wisdom. We can’t time
the cycle, and we can’t ignore the consensus.


Normalized Earnings
We can construct a kind of parallel valuation universe based not on specifi-
cally timed events but on the average outcome of the inevitable. In other
words, we can “normalize” revenues and, most notably, earnings. Our for-
mal income growth models are based mainly on the things a company can
control—its costs, its technology, its best practices and processes—and only
somewhat on the unpredictable. We put these assumptions out there,
because to do otherwise—to bake in our own recession, for example—
would be to create a misleading variant for clients who are trading on our
advice right now.
      But for many of our models, we separately and privately track valuation
metrics based on normalized earnings. As the sketchy clouds of a spent-up cycle
coalesce into the gray cumulus of oncoming recession, the normalized earnings
trend assumes a larger place in our thinking. And most often, the growth rates
informing our normalized EPS projections into the future are based on our OLS
growth calculations.
      In summary, OLS is based on a smoothing of historical data, permitting a
more comprehensive look-back to help predict the future. By minimizing the
importance of aberrant CAGR beginning or end points, it provides a smoothed
growth rate that is useful in minimizing or eliminating temporary company
strategy effects while capturing economic cycle effects. It can overcome the limi-
tations of negative beginning or end points. OLS is also a broad brush that can
be applied to growth in any metric, from pro forma EPS to how many times a
year you change your shoelaces.
      Normalized earnings are more forward looking and, as the name implies, are
of most use in the realm of profits projection (although we could hypothetically
normalize the shoelace changing as well). To my knowledge, any OLS calculation
is going to look almost the same. There are, however, numerous ways to normalize
forward earnings; we’ll discuss a few and then indicate our preferred
methodology.
      As the previous discussion suggests, it makes sense to calculate OLS growth
rates based on historical performance so we have them available for use in our
                   Ordinary Least Squares Regressions and Normalized Earnings   •   125



normalized earnings calculations. On that basis, we’ll first describe the OLS pro-
cess; we’ll then move on to a discussion of normalized earnings (NE).
      Also note that we are discussing OLS and normalized earnings in a single
chapter, and we are going to situate the OLS calculations and our normalized
earnings work on a single worksheet. This simply reflects the fact that we recon-
figured a “normal” normalized worksheet from vertical by years to horizontal by
years. Feel free to use separate worksheets, if you prefer.


OLS Growth Rates
Worksheet Setup and Method
The OLS and NE worksheet uses annual data ranked left to right. Essential to this
process and the OLS process is the inclusion of all years in our measurement. In
this case, our example shows the years 1996 to 2008. You can work with much
shorter or longer periods as needed.
      We can link to much of this data, conveniently, from our ratios and valua-
tion worksheet, where most of the standard financial statement data (i.e., income
statement, balance sheet, and cash flow statement) resides in annual stacks. Even-
tually you’ll determine what is most useful in your analysis. Here are the basics
you would normally want to include:

     From the Income Statement Presentation:
     • Revenue
     • Operating income (GAAP or pro forma)
     • Pretax income (GAAP or pro forma)
     • Net income (GAAP or pro forma)
     • Per-share earnings (GAAP or pro forma)
     • Any segment revenues and operating income

     From the Balance Sheet:
     • All-in cash
     • Accounts receivable and inventories
     • Accounts payable
     • All-in debt
     • Total assets
     • Total liabilities
     • Stockholders’ equity
126   •   Income Statement Presentation



      From the Cash Flow Statement:
      • Depreciation and amortization
      • Net working capital cash flow (use)
      • Cash flow from operations
      • Capital expenditures
      • Dividends

      This is one worksheet that, if properly constructed, can generate data rather
quickly, because the calculation strings are designed for easy copy and paste. Our
example company in this exercise is Motorola (MOT), an industry pioneer with
an unmatched legacy that has had difficulty executing in the modern age. We
chose this troubled giant, not out of sentiment, but because companies whose
performance oscillates wildly and swings from profits to loss normally frustrate
the simple calculation of growth rates.
      If we look at Motorola’s revenue line, we do not see much progress. From
1996, when sales were $27.97 billion, to 2008, when revenues were $30.15 billion,
revenues grew just 0.5% annually, according to our simple end point–to–end
point CAGR calculation.
      The standard CAGR formula we use to get this calculation is (N5/
B5)^(1/13) 1), where N is 2008 revenues and B is 1996 revenues. To get
to the appropriate discount rate, we use the “^(1/13)-1” portion of the
formula to discount the growth over the appropriate amount of years, in
this case 13.
      To calculate an OLS growth rate for this period, we need to derive OLS
values for all the years 1996 to 2008; the string of years, in turn, will normalize
the beginning and end points. If the year 1996 is in cell B3 and actual revenues
are in cell B5, the formula in cell B6 is TREND($B5:$N5,$B$3:$N$3,B$3).
Drag and drop this formula across the page to cell N6, directly beneath actual
revenues for 2008. Figure 7.1 is a first look at an OLS worksheet; it shows an OLS
regression for Motorola’s revenues from 1996 to 2008, as well as a five-year OLS
regression for 2004 to 2008. (To make the model fit better on the page, we have
hidden years 1998 to 2001; but they are part of the formula.)
      In no other formula we use is it so vital to get the dollar signs right! Taking
apart the formula TREND is what it says it is, providing us the underlying trend
across this specified set of data points. In the first part of the formula, the dollar
signs before B5 and N5 make sure you keep the range intact (1996 revenue to
2008 revenue) across all data points. The second range cites the years and uses
dollar signs to lock both the left-to-right range (the years 1996 to 2008) and to
ensure that the years are maintained as you go down the page. The third part of
Figure 7.1

Motorola’s revenues have been on a wild ride, ranging from $26 billion to $42 billion. If we smooth the growth rates with OLS rather than relying solely on the end
points for our CAGR, we see that annual growth averaged closer to 1.5% than to the 0.6% suggested by the unadjusted end points.

Motorola
Ordinary Least Squares
                 1996           1997              2002        2003         2004         2005        2006         2007         2008
Incm Stmnt
Revenue         27,973.00      29,794.00         26,681.00    23,155.00    31,323.00   36,843.00    42,879.00    36,622.00    30,146.00    0.58%      Unadjusted
                28698.011      29224.214         31855.231   32381.434    32907.637    33433.841   33960.044    34486.247    35012.451     1.54%   OLS Smoothed
                                                               5-Year:     31,323.00   36,843.00    42,879.00    36,622.00    30,146.00   -0.76%
                                                               5-Year:       36077.6     35820.1     35562.6       35305.1     35047.6    -0.58%
128   •   Income Statement Presentation



the formula allows the individual year to move laterally but keep the formula
linked to the year line.
      Now that we have smoothed values for all periods, we can use our standard
CAGR formula to derive a “truer” growth rate. Using the (N6/B6)^(1/13) 1
formula, our OLS-adjusted CAGR tells us that Motorola’s annual revenue growth
across this 13-year span was actually closer to 1.5% annually.


Reading the Data: Income Statement
Even though they can be volatile, revenues are one of the relatively steadier line
items in most income statements. And, notwithstanding a few biotech companies
(most of which I’ve owned at one time or another), it is hard to have negative
revenues.
      As companies move down the income statement—particularly mature
giants like Motorola, constantly swatting at the upstarts continually upsetting
their markets—producing consistent profit growth is an endless challenge. You
need go no further than the gross margin line to see where this company’s trou-
bles begin. Although OLS “smoothed” revenues grew at a 1.5% annual clip from
1996 to 2008, cost of goods sold increased a smoothed 3% annually. Conse-
quently, gross margin has declined a smoothed 1.1% annually over this span—
double the 0.5% rate that a simple CAGR calculation produces.
      If we want to measure Motorola’s operating income over this span, we
immediately confront the GAAP operating loss of $2.39 billion for 2008. Anyone
who has used a CAGR point-to-point formula in the past knows that negative
values in either of the end points whacks out the formula and makes its useless.
According to a simple CAGR calculation, Motorola’s operating income has
declined 200%—every year. Even this company is not that troubled.
      Let’s now adjust the GAAP operating income using an OLS regression; the
regression is shown in Figure 7.2. The regression has the effect of cutting 1996
operating income by one half, to $1.25 billion from an actual $1.96 billion. But
the regression also produces a positive end point. The CAGR formula works with
the positive beginning points and end points, showing us that Motorola’s GAAP
net income has declined 5% annually since 1996. On the other hand, over the
2004–2008 span the operating profit decline was so severe that even OLS cannot
produce a positive end point, thus rendering the CAGR calculation
meaningless.
      The story is the same in pretax income and net income. GAAP pretax
income is down over 200% annually from 1996 to 2008 according to the useless
simple CAGR formula; using an OLS regression, pretax income edged up one half
Figure 7.2

Revenues will be a much steadier input than operating income for almost any company, and the likelihood of operating losses in any one year is greater. OLS can
smooth away negative end points over long spans, although it cannot work miracles amid deep declines in profits.

Motorola
Ordinary Least Squares
                               1996            2002         2003        2004        2005        2006         2007        2008

 Oprtng Incm                   1,960.00       (1,750.00)    1,273.00     3,150.00   4,696.00    4,092.00     (553.00)   (2,391.00)   -201.54%    Unadjusted
                               1247.537          932.413    879.893      827.372     774.852     722.331      669.810      617.290     -5.27% OLS Smoothed
                                                            5-Year:      3,150.00   4,696.00    4,092.00     (553.00)   (2,391.00)   -194.64%
                                                            5-Year:     5065.000    3431.900    1798.800      165.700    -1467.400   -178.05%
130   •   Income Statement Presentation



a percent annually. Net income, down 210% on simple CAGR, is off a more rea-
sonable, though hardly pretty, 9% annually since 1996.
      Use of OLS regressions on pro forma results can get an analyst in trouble, and
the Motorola example shows us why. While a simple CAGR calculation shows a
175% annual decline in pro forma net income, we calculate an 11% pro forma net
income growth rate based on OLS. If you were unaware of the laundry list of prob-
lems facing Motorola, you might think this growth rate can be blithely plugged into
formula. We are not (ever) passing judgment on a mathematically sound formula,
but be aware that use of OLS regression on pro forma anything amounts to a double
cooking of the books and should be used with eyes wide open.
      Is OLS regression of end points always superior to simple CAGR calcula-
tions? We can think of one example where the answer is no. In 2008, Motorola’s
enterprise value was $35.9 billion; at the end of 2008, it was closer to $23 billion.
A smoothed regression of this change suggests a 2% annual decline, whereas the
actual point-to-point decline was 3.3% annually. We are not looking to use the
OLS data to model change in enterprise value; we are more interested in whether
the decline signals a good entry point in the stock or a bad one. To make this
decision, we need the actual numbers. The 3% actual point-to-point decline in
enterprise value is more favorable to the bull-value case than the 2% OLS-derived
decline.
      Moving back to revenues and revisiting Figure 7.1, we now look at the com-
pany’s simple CAGR and OLS-smoothed CAGR over the market’s look-back period,
which tends to be five years. Without delving too deeply, we note that this has been
a tough patch for Motorola, as its global mobile handset share was cut to one third
of peak levels, and its wireless networks business contracted meaningfully. Pricing
became more difficult, the business mix worsened, and cash—depleted by asset
buys (of Symbol Technologies)—was reduced as an earning asset.
      Accordingly, while five-year revenue growth was declining 1 point annually,
gross margin was declining at five times that rate. For the preceding five years,
the GAAP end points for operating income, pretax income, and net income are
negative, resulting in screwy negative growth rates, even on an OLS basis. On a
pro forma basis, the end points are positive, making it possible to calculate mean-
ingful growth rates; however, they’re not good, typically declining in double dig-
its over the preceding five years.


Reading the Data: Balance Sheet
The balance sheet has shown its strains as well. As illustrated in Figure 7.3, over
the long time frame, cash grew much faster (11% annually) than debt (less than
Figure 7.3

In addition to difficulty with operating profits, Motorola’s net was impacted in recent years by an accelerating decline in cash matched with a rising pace of
debt growth.

Motorola
Ordinary Least Squares
Balance
Sheet      1996     1997        1998      1999      2000       2001       2002      2003      2004      2005      2006            2007     2008
 Cash     3,509.00 3,442.00     3,451.00 5,536.00   5,301.00   9,116.00   8,619.00 11,351.00 13,949.00 15,208.00 16,535.00       9,443.00 7,496.00      6.01% Unadjusted
 Invstmnts
 Total
             3585.297 4435.901 5286.505   6137.110 6987.714 7838.319 8688.923 9539.527 10390.132 11240.736 12091.341 12941.945 13792.549               10.92% OLS
                                                                                                                                                               Smoothed
                                                                                     5-Year: 13,949.00 15,208.00 16,535.00 9,443.00 7,496.00           -11.68% Unadjusted
                                                                                     5-Year: 16260.400 14393.300 12526.200 10659.100    8792           -11.57% OLS
                                                                                                                                                                Smoothed

 Debt        3,313.00 3,426.00 5,542.00 5,593.00               9,242.00   8,818.00   7,571.00   5,295.00   4,319.00   4,397.00   4,323.00   4,184.00    1.81%
 Total                                           10,684.00
             6267.791 6206.582 6145.374 6084.165 6022.956 5961.747 5900.538 5839.330            5778.121   5716.912 5655.703 5594.495 5533.286         -0.95%
                                                                             5-Year:            5,295.00   4,319.00 4,397.00 4,323.00 4,184.00         -4.60%
                                                                             5-Year:              4947.2     4725.4   4503.6   4281.8    4060          -3.88%
132   •   Income Statement Presentation



1%). In the preceding five years, however, cash declined faster than debt. Analysts
do not operate in an information vacuum, so the changes need to be seen in
context. Subjectively, we feel the $4 billion spent to buy Symbol Technologies was
better than the $2 billion spent on share buybacks.
      On the upside, MOT likely hit bottom at the cycle bottom—and this was a
deep cycle bottom. The difficulties in the handset business have distracted inves-
tors from incredible franchises in connected home and first-responder networks.
Other areas are benefiting from the reservoir of engineering excellence, but this
cash-rich company is running leaner. For these reasons, it may be worthwhile to
calculate Motorola’s normalized earnings, in order to get some sense of the com-
pany’s potential multiyear earnings power.


Normalized Earnings
NE Worksheet Concept
Don’t confuse normalized EPS projections with the real thing, particularly on a
single-year basis. Normalized earnings mainly have value in projecting a com-
pany’s potential earnings power across a bigger span: the economic cycle, for
example, and perhaps the years leading in and out.
      What does it mean to “normalize” earnings? The process involves calculating
historical growth rates and margin percentages and using them to project earnings
out over a multiyear period. While OLS calculations are usually done one way, the
same is not true for normalized earnings. Given the many significant line items in
the income statement, it is possible to trend any combination of them to arrive at
normalized earnings. The lack of a single standard makes them all the more awk-
ward to implement in a model, even a “shadow” model.
      Most investors agree that normalizing earnings means to adjust them to
eliminate variations related to the business cycle. Many approaches are backward
looking only. That is, normalized earnings are nothing more than the simple
average earnings for a given period. Given the rapid shifts in the market, that
kind of calculation is not much use in attempting to determine future earnings
power.
      Our NE method is a hybrid, which is the modern way of saying “ram-
shackle.” Despite its less than blue-blood origins, we think it hits all the main
points, or rather draws from the appropriate data, to derive an estimate of the
average outcome.
                   Ordinary Least Squares Regressions and Normalized Earnings   •   133



Normalized Earnings Method
Stage One: NE Compilation
As noted, our normalized earnings process is sited on the bottom of the OLS
worksheet so we have handy access to that data. Start by referring to Figure 7.4 to
get some sense of the process. This is effectively a two-stage process. In stage one,
we compile data and derive margins on that data; in stage two, we extrapolate the
data growth trends and rolling five-year margin trends out a few years to derive
EPS estimates.
      While we consider this to be a two-stage process, note that data for stage
one in the Motorola model has been compiled on both an unadjusted basis and
also on a smoothed or OLS basis. (For both we used pro forma inputs, not to
deliberately add to the confusion but because the market values MOT that way.)
If you have a healthy company with somewhat steady top-line growth and margin
expansion, you can forgo use of OLS values in your normalized earnings
calculation.
      In stage one, we “stage” key inputs linked from our actual and OLS-
smoothed data above and then use margin data implied by these figures to cal-
culate EPS afresh. In the first part of stage one, we show the non-OLS inputs, and
for this company they end with a thud: our 2008 EPS is a one-penny loss. This is
not a good basis to go forward; so we rely on the second set of stage one inputs,
which uses OLS-based inputs.
      In this second rendering of stage one and using smoothed data across a
13-year span, we see that operating margin moves in a tighter 4.1% to 1.8% range.
Net margin is actually higher (reflecting all those years when the cash horde was
throwing off huge interest income). We arrive at nicely positive “smoothed”
earnings of $0.50 by 2008.
      In the third part of stage one, where we’ve compiled the OLS tallies (in
Figures 7.1 through 7.3 these were shown on the far right), we see a relatively
healthy 9% growth rate for pro forma earnings. Remember, keep the eyes wide
open when using both OLS and pro forma in the same calculation.


Stage Two: NE Application
In stage two, we apply the revenue, income, and margin data we derived in stage
one. We begin with at least five preceding years, so we have a place to compile
five-year average margins. In the first line, we begin with actual OLS-derived
Figure 7.4

Normalizing earnings can help paint a picture of a company’s average EPS strength across coming periods. When this data is built on OLS inputs, in our view, it
may better capture performance across the ups and downs of the earnings cycle.

Motorola
Normalized Earnings

 Stage ONE: NonNrml
 w/o OLS             1996       2001       2002     2003     2004      2005      2006      2007      2008
 Revenue            27,973.00 29,451.00 26,681.00 23,155.00 31,323.00 36,843.00 42,879.00 36,622.00 30,146.00
 Pro Forma Op Incm   1,960.00 (636.00)       847.00 1,273.00 3,150.00 4,696.00 4,286.28      642.00    373.74
 Oprtng Margin           7.0%     -2.2%       3.2%     5.5%     10.1%    12.7%     10.0%      1.8%       1.2%
 Pro forma Income    1,154.00 (3,937.00) (2,301.00)   932.00 2,199.00 4,599.00 3,186.67      569.15   (30.26)
 Net Margin             4.1%     -13.4%      -8.6%     4.0%      7.0%    12.5%       7.4%     1.6%      -0.1%
 Shares Out Diluted  1,836.60   2,212.75 2,282.00 2,351.20 2,466.20 2,521.16 2,506.37 2,341.14 2,266.60
 Pro forma EPS            0.63     (1.78)    (1.01)     0.40      0.89     1.82       1.27     0.24     (0.01)

 Stage ONE: NonNrml
 w/OLS               1996            2001      2002      2003      2004      2005       2006     2007      2008
 Revenue            28,698.01       31,329.03 31,855.23 32,381.43 32,907.64 33,433.84 33,960.04 34,486.25 35,012.45
 Pro Forma Op Incm   1,247.54          984.93    932.41   879.89     827.37    774.85    722.33    669.81    617.29
 Oprtng Margin          4.3%            3.1%      2.9%      2.7%      2.5%      2.3%       2.1%     1.9%      1.8%
 Pro forma Income      319.45          765.10    854.23   943.37 1,032.50 1,121.63 1,210.76 1,299.89 1,389.02
 Net Margin              1.1%           2.4%      2.7%      2.9%      3.1%      3.4%      3.6%      3.8%      4.0%
 Shares Out Diluted  1,970.06        2,198.44 2,244.12 2,289.79 2,335.47 2,381.14 2,426.82 2,472.49 2,518.17
 Pro forma EPS            0.16           0.35      0.38      0.41      0.44      0.47       0.50     0.53      0.55
Figure 7.4 (continued)


Motorola
Normalized Earnings

                             1996     2001      2002       2003      2004      2005      2006       2007      2008
 OLS Adjusted:
 Revenue CAGR                 1.54%
 PF Op Incm CAGR             -5.27%
 Aveage Oprtng Margin         2.98%
 PF Net Incm CAGR            11.97%
 Average PF Net Margin        2.63%
 Diluted Shares CAGR          1.91%
 PF EPS CAGR                  9.88%

 Stage TWO:
 Normalized w/OLS                      2001      2002      2003      2004      2005      2006      2007      2008      2009E     2010E     2011E     2012E
 Revenue                              31,811.99 32,302.39 32,800.36 33,306.00 33,819.44 34,340.79 34,870.18 35,407.73 35,953.57 36,507.82 37,070.62 37,642.09
 Oprtng Incm                            933.04     883.89    837.32    793.21    751.42    711.83   674.33    638.80     605.15   573.27     543.07    514.46
 Rolling 5-year Net Margin                1.7%      1.9%      2.2%      2.4%      2.7%      2.9%      3.1%      3.3%      2.9%      3.0%      3.1%       3.1%
 Pro forma Income                        529.16    623.18    716.98    810.66   904.27     997.90 1,091.61 1,185.46 1,042.09 1,092.07 1,132.45        1,161.12
 Shares Out Diluted                    2,198.44 2,244.12 2,289.79 2,335.47 2,381.14 2,426.82 2,472.49 2,518.17 2,566.17 2,615.08 2,664.93 2,715.73
 Pro forma EPS                             0.24      0.28      0.31      0.35      0.38      0.41      0.44      0.47      0.41      0.42      0.42       0.43
136   •   Income Statement Presentation



revenue from 2001 through 2008, and then grow revenue for forecast periods
(2009 through 2012) at the OLS CAGR of 1.54%. Our formula would be some-
thing like N201*(1 $B189), where N201 is 2008 OLS revenue and B189 is the
1.54% growth rate. (Remember, we do this work way down at the bottom of an
actual OLS worksheet. In the snippet shown in Figure 7.4, 2008 OLS revenue is
found in cell N40, and the 1.54% growth rate is found in cell B28.) Although we
include an operating income line, our main driver will be net income. So we
average five-year net margin (again based on the OLS calculations) and apply the
rolling average beginning in 2009. The rolling five-year net margin applied to
revenue drives our net income forecast for the forward years.
      We also grow out the share base at the OLS growth rate, and that finally
provides a basis for our modeled pro forma EPS for the years 2009 through 2012.
In the case of Motorola, our model rendered fairly consistent EPS across the
period. This is a mechanical process based on the data, not our perceptions and
subjective assessments of the company’s progress in a dynamic market. Motorola
has shown Houdini-like skill in extricating itself from bad operating models.
With the admirable Sanjay Jha at the company’s helm, we have confidence it will
beat our normalized outlook. But if the economy comes crashing down again,
these numbers may be the more reliable figures for modeling the company.
      Companies with steadier sales growth and margin expansion will typically
produce rising normalized earnings, even without the assist of pro forma data
and OLS-adjusted inputs. But we prefer to use OLS inputs when we normalize
even the healthiest companies, because this in our view best captures past eco-
nomic cycles and pushes that data forward into our NE calculation.
      Granted, we can’t fully “normalize out” the cycle even for the best-function-
ing companies. Over the 2007–2009 span, we learned that fully informed reces-
sions (those that have real-time information on demand, inventories, and all
other inputs) hit harder and faster than the old, blind recessions that used to
stumble forward in a knowledge vacuum. Hopefully, informed recessions lead to
quicker recoveries; that’s what the inventory software salespeople promised.
Given this quick-change reality, normalized earnings capacity could take on
increased importance within the modeling architecture.
      With OLS and normalized earnings now in the tool kit, we really are wind-
ing down the modeling-intensive portion of our work. It is now time to consider
the next part of the analyst’s task: to take our modeled or forecast inputs, inte-
grate them with historical data, and, based on their interaction, to begin valuing
the equity.
                                                             PART 2

           RATIO AND
           VALUATION
           WORKSHEET



Introduction
In this second part of the book, our objective is to compile annual data from the
main financial statements to perform ratio analysis. We’ll use historical and
modeled inputs from the financial statements as well as historical, real-time, and
forecast asset prices to estimate asset value based on comparable historical analy-
sis. And we’ll lay the groundwork for present value calculations to be performed
on a separate worksheet.
      Now that we’ve built our income statement projection, we need to gather
historical data from two other key financial statements: the balance sheet and the
income statement. We’ll also need to project these financial statements out for
our modeling period of two years. For the purpose of compiling ratios and per-
forming historical comparables analysis, at a minimum we need to model annual
totals for the balance sheet and cash flow statement projected out at least two
years.
      We can model the balance sheet and cash flow statement with real rigor, so
that the accounts in each sheet are interleaved with one another. Certainly, for
the sell-side analyst in training, that level of modeling is expected. At the same
time, it can be efficacious (better known as “quick and dirty”) to model growth
in balance sheet accounts in line with historical experience, typically as some
                                                                                137
138   •   Ratio and Valuation Worksheet



multiple of GDP growth or in relation to the company’s own revenue growth.
And we can model a cooked-down cash flow statement that mainly focuses on
the key accounts that will drive our discounted free cash flow model.
      Most sell-side analysts keep a running quarterly balance sheet, either at the
very bottom of the income statement presentation or on a separate worksheet
(which is our usual practice). Some keep a running quarterly cash flow statement
as well (we’re usually remiss). But on this worksheet, we will focus primarily on
annual balance sheet and cash flow assumptions, because historical comparable
valuations such as price/earnings ratios, or P/Es, are normally expressed on an
annual basis.
      The ratio and valuation worksheet meets several goals. It is a place to com-
pile common ratios that give us some sense of a company’s performance in sev-
eral broad areas. These include internal liquidity, operating efficiency, operating
profitability, risk analysis and profile, and profits and cash flow. Some of that
data figure immediately in our comparables valuation analysis. Some of it will
figure in our discounted free cash flow analysis. Further out, much of it will be
linked to a matrix or master spreadsheet on which we compile and compare data
from multiple companies. And finally, the collated information on the matrix
spreadsheet will be manipulated to produce a proprietary peer-determined value
(PDV) that quantifies a company’s value relationship within a designated peer
group. This proprietary metric is one of several that will ultimately figure in our
final assessment of the dollar value of the equity.


Historical Comparables and DFCF: Is It Either/Or?
In the third section of the book, we will prepare a worksheet for discounted free
cash flow (DFCF) analysis. This form of analysis projects (and discounts) future
cash flows and also calculates (and discounts) a “terminal” value for the asset.
Accordingly, it is regarded as mainly forward looking. While DFCF is predicated
on a historical performance across a time frame, that period is typically not long,
on average two to three years. DFCF, with its summed future cash flows and
terminal value, all discounted to the present, most readily lends itself to creation
of a dollar value for the asset.
      Comparable historical valuation, on the other hand, uses a longer historical
period (typically five years) to model the valuation outlook for a relatively short
span—say two years. Use of a five-year historical period is designed to encompass
all phases of the economic cycle as a means to guard against cyclical distortions.
Given the shorter forward time span, discounting back to present value is not
                                                 Ratio and Valuation Worksheet    •   139



regarded as necessary in comparables analysis. Comparables valuation also gen-
erates a like value—for example, historical P/E is used to determine a likely for-
ward range of P/E—whereas DFCF is designed to predict or calculate a dollar
value for the asset. Historical comparables data, however, can be used to model
a dollar value for an asset, and we’ll share our methodology for doing so
presently.
      The two valuation styles—historical comparables and DFCF—move in and
out of vogue with analysts, much as dividend-paying and non-dividend-paying
stocks move in and out of vogue with investors. To reiterate our approach: we’re
not here to argue relative merits of forward-looking versus backward-looking
valuation. We’ll take their merits as a given and use them to help determine dol-
lar value of an asset. What we will do in our final analysis is assign each style (and
a third peer-based valuation style as well) a weighting within our final determi-
nation of dollar value of an asset, based on the market climate and other objective
and subjective considerations.


Historical Comparables: A Primer
The heart of comparables valuation is determining a basis of historical valuation
relationships that will be used in tandem with modeled inputs to forecast future
valuation measures. These valuation measures are typically based on the inter-
play between the price of an individual common share and various per-share
metrics, the most common of which are price to earnings, price to revenue, price
to book value, price to cash flow, and relative P/E.


Overreliance on P/Es
Our goal is to determine what the stock is worth on a dollar basis, adjusted for its
riskiness within its trading universe. Equity analysis has been described as an art and
a science. That is a fair surmise; unfortunately, it is also describes a gray zone where
valuation analysis can grow fuzzy and—most often—fall back on the familiar.
      Anyone who has earned an MBA in finance or who has been through the
CFA program has encountered a plethora of valuation methodologies. These
sometimes follow the fashion of the moment; the long-disparaged dividend dis-
count model (DDM) seems to be cycling back into favor, as it often does when
growth stocks fall from the heavens. In the academic finance setting, valuation
methodologies such as DDM, discounted free cash flow (DCFC), and historical
comparable methodologies all are given a fair schooling.
140   •   Ratio and Valuation Worksheet



       Ultimately, analysts are left on their own to conclude how much weight to
assign to any valuation methodology. With multiple methodologies percolating
in their heads, newly hatched analysts are plunked down in the real world. Men-
tors notwithstanding, these new analysts must sort for themselves to find which
methods best fit the available information set. At a micro level, the markets and
stocks respond to real-time news flow. Not surprisingly, analysts eventually grav-
itate to the valuation methodology that encompasses the most current and plenti-
ful information. This information can come from all sources, ranging from
industry blogs and boards to chats with IR people and other analysts, or breaking
news in the Wall Street Journal or the Financial Times.
       Often, breaking information will directly or indirectly impact earnings.
What to make of the news that Ceylon Telephone & Telegraph is coping with
diminishing pay-phone revenues? It is not easy to reflect such minutiae in a DDM
or DCFC model. Such models are meant to determine a long-range or ultimate
value. Clients don’t want to hear that, of course, when their Ceylon Tel & Tel
shares are getting killed.
       But every analyst knows that they can back out some portion of the Borneo
pay-phone contribution from projected earnings; voilà, there’s an instant change
in the forward P/E. Moreover, because the analyst has been digesting information
and adjusting earnings all through a given period, he or she can track the resul-
tant changes in the EPS trend and—in league with an assessment of the stock
price trend—monitor the resultant trend in P/Es.
       And that is why analysts, from the most tightly focused bulge-bracket spe-
cialist to a buy-side generalist charged with covering a few sectors, rely—overrely,
really—on P/Es. There’s another reason why analysts fall back on P/Es: it’s what
the investing public wants. Consider the individual charged with managing
someone else’s money. Whether a fresh-faced financial consultant or an ensconced
private wealth manager, he or she is expected to be conversant in the 5,000 stocks
trading on the NYSE and the Nasdaq main exchange. The Street cries out for a
tabla rasa, a common metric that reduces a company’s worth of financial com-
plexity into a simple ratio; P/E is our Rosetta Stone.


The Timing Problem with P/Es
Earnings are an accounting fiction, a witch’s brew of data influenced by thou-
sands of sub-sub-segment P&L decisions made regarding revenue recognition,
accounting for inventories, depreciation schedules, research and development
                                                Ratio and Valuation Worksheet   •   141



(R&D) cost assignment, compensation timing, tax assessments—the list goes on
and on. There is surprising leeway in many of these decisions, regardless of
whether or not a company is tempted to make the most favorable interpretation.
On a good day, earnings are suspect.
      The problem with P/Es in isolation is not that that they are based on this
fiction. Any metric, even one predicated on false data, can have useful informa-
tion content as long as it is universally applied. The problem with P/Es is that they
provide such a limited amount of information and are so easy to misread at dif-
ferent times in the economic cycle.
      The investor with a little bit of knowledge senses that stocks should be
bought when P/Es are low and sold when P/Es are high. But everyone in the busi-
ness knows that highly cyclical stocks should be bought when P/Es are high and
sold when P/Es are low. That strategy certainly worked in the spring and summer
of 2009, when the most money was being made. This is counterintuitive, and—
back to our wealth manager with 5,000 tickers dancing before his or her eyes—
that knowledge can get lost when in so many other areas, low P/Es are the right
prescription.


A More Balanced Approach to Comparables Valuation
Earnings may be no more tangible than a hallucination, but at least they are a
shared hallucination. Because every company determines earnings in more or
less the same way, this creates a common ground not just for measuring growth
and margins but for valuing the asset.
      Our goal, however, is to use not just earnings but the breadth of income
statement, cash flow statement, and balance sheet data to derive a more balanced
picture of the ongoing enterprise. Moreover, we are not content to just determine
a ratio and see how it “feels” when we lick our finger and hold it up in the air. As
always, our goal is to use information—in this case, the ratio and historical infor-
mation that contributes to what we call historical comparable valuation—to
derive a fair value assessment of the asset in dollars. As much as possible, we’re
going to dispense with the rubbery ranges that sometimes strip value from these
assessments. We want a hard dollar value.
      As often as not, that hard dollar value will be wrong. But in the process of
creating hard dollar values for assets and then tracking their variance from
actual value over a period of years, we will refine the calibration until we get it
right.
142   •   Ratio and Valuation Worksheet



The Ratios and Valuation Worksheet Described
We have been meticulous in modeling income statements on a quarterly basis,
because this level of flexibility and granularity best mirrors the market’s demand
that we incorporate the timeliest data. Ultimately, however, the most common
valuation metrics—absolute and relative P/Es, discounted free cash flow valua-
tions—are based on annual assessments of a set of inputs, be they earnings, cash
flow from operations, or another data point. Again, along our cattle drive, we’re
not going to stop to question this valuation convention, any more than we stop
to question the ethics of branding (as opposed to, say, matching sweaters for the
cattle). Our job is to keep those doggies rollin’.
      We need a place to “stack up” annual data so it can be measured, manipu-
lated, modeled, and used in ratio and valuation analysis. Our ratios and valua-
tions page is where lots of this information will be stored and then used to assess
valuation according to conventional and traditional metrics.
      The top 40 to 60 lines of the ratios and valuation worksheet will be devoted
to common ratios. We will then dedicate a roughly comparable space to valuation
analysis based on historical comparables: P/E, the price/sales ratio, the price/
book value ratio, the price/cash flow ratio, the price/ free cash flow ratio, and
relative P/E. The data that forms the basis for valuation analysis—primarily
annual compilations of the income statement, balance sheet, and cash flow state-
ment—will be located on the bottom of the worksheet.
      The ratios and valuation worksheet has five sections:

      •   Cash flow statements
      •   Annual financial income statements
      •   Balance sheets
      •   Ratios
      •   Historical comparable valuations

      Because we need the financial statement data first, we’ll work our way down
the previous bulleted list. As such, cash flow statements, annual financial income
statements, balance sheets, and ratios will be discussed in Chapter 8 and Chapter
9 of Part 2 of this book, while historical comparable valuation will be covered in
Chapter 10.
                                                  Chapter 8
           RATIO ANALYSIS, PART 1:
           INTERNAL LIQUIDIT Y AND
           OPERATING EFFICIENCY




Annual Financial Statements
We cannot perform ratio analysis and valuation analysis without data, so we
begin with annual data that is either imported or added directly on the page. We
will be importing from, or more specifically linking to, our income statement
presentation to get the annual tallies. But what about the balance sheet and the
cash flow statement?
      The answer lies with your work assignment and the depth of coverage you
can bring to a subset of individual equities. The sell-side analyst is tasked typi-
cally with analysis of 12 to 25 companies in a tightly defined universe. This
analyst will be visiting management, participating in quarterly conference calls,
and generally modeling his or her covered companies with a high level of detail.
Moreover, high-velocity traders such as hedge funds that offer the highest com-
missions will expect these analysts to have the most current analysis, modeling
almost in real time. The sell-side analyst needs access to all the publicly avail-
able information to best serve these real-time needs; equally important, this
relatively small universe needs to be meticulously modeled. A private wealth
manager working in a tightly defined style—say, very large cap blue chips from
the Dow 30 or the S&P 100—may also have the time and resources to model at
this level.
                                                                               143
144   •   Ratio and Valuation Worksheet



      Accordingly, these analysts and wealth managers will typically maintain
quarterly balance sheets and quarterly cash flow statements as separate work-
sheets within the workbook, or alternately far down the income statement. Much
as we extract only the annual summary columns on our income statement for
linkage to the ratios and valuation (R&V) page, the specialty analyst or wealth
manager will link the year-end columns from these individual worksheets to the
R&V worksheet.
      Portfolio managers and analysts charged with monitoring an entire sector,
managers whose portfolios are driven more by style (e.g., growth or value) than
by sector, buy-side analysts, and generalists in general find it possible to dispense
with maintaining separate worksheets for quarterly balance sheet and cash flow
statement data. This group, in their income statement presentations, likely did
not have the luxury of “going granular” to the workbench level, and likely mod-
eled the P&L only and perhaps any available segment data. For these generalists,
inputting annual balance sheet data and cash flow statement data directly onto
the page is sufficient; so too is modeling these financial statements at an annual
rather than quarterly level.
      Given that valuation analysis is based on annual data inputs, there is no
difference in the outcomes when building valuation “bedrock” based on already
reported (historical) periods whether data is originated on a supporting quarterly
worksheet or copied directly onto the R&V worksheet. In a bit, we’ll get to model-
ing annual balance sheet and cash flow statements. Here, too, the scope of the
analysts’ duties will influence the sophistication of the modeling process; we’ll
offer various options.


Practical Tips in Data Inputting and Linking
To reiterate, our work is not intended as an Excel primer; that topic has a groan-
ing bookshelf of its own, and we can’t possibly cover even a smattering of key
points. But we’ll share some practical tips and observations based on what we’ve
discovered in building R&V worksheets.
      At its most basic, linking a cell from one worksheet to another requires the
user to type the equal sign in the target cell, switch to the worksheet, and click
on the individual cell where the data resides, and affirm the link either by press-
ing Enter or by clicking the checkmark cell in the task pane.
      When a cell is linked from another worksheet within the workbook, you can
drag and drop a stream of linked data. This makes its easy to import and link
data from the income statement presentation worksheet or from quarterly bal-
ance sheet and cash flow statement worksheets (if maintained). When attempting
              Ratio Analysis, Part 1: Internal Liquidity and Operating Efficiency   •   145



to link from a different workbook, however, keep in mind that the linked cell will
include the $ symbol in front of the row and column designation. If you attempt
to drag and drop a cell with this designation, you will instead replicate the source
cell again and again. So when linking from a different workbook, as a first step
remove the $ symbol within the cell.
       In linking to the income statement presentation worksheet, begin with the
line item headers (e.g., revenues, R&D, interest income) usually in column A.
Drag and drop the link to encompass everything from the period designation
(e.g., 1997) to the dividend. We do not need to link and copy the line items that
we use to refine our quarterly model (e.g., cash and debt) or distinguish GAAP
from pro forma calculations (e.g., FAS 123R and amortization).
       If our line item headers on the income statement presentation worksheet are
in column A and our first historical year (say, 1997) is in column B, on our R&V
worksheet we can drag and drop the entire 1997 historical year in a single step.
Simply highlight the line item headers down to dividend and drag and drop them
one column to the right. As soon as we get to our first quarterly breakdown on
the income statement presentation—say, 2002—we need to change the link to
reflect the change; otherwise we’ll be dragging and dropping 1Q02 instead of full
year 2002.


Populating the R&V Worksheet with Annual Financial Statements
As a practical matter, we find that the annual financial statement stacks start
between row 80 and row 85 on the R&V worksheet. We also note that we consis-
tently order our financial statement presentations with the income statement on
top, followed by the balance sheet and then the cash flow statement. We do not
typically include a model of the financial statement for the changes in stockhold-
ers’ equity.
      The easiest way to populate the annual income statement portion of R&V
is to link each annual revenue cell from the income statement presentation,
arrange them all next to one another, and drag and drop down the page. (Unlike
on the income statement presentation, we do not need adjacent columns for
annual or sequential comparisons.) In addition to the historical period annual
income statements, we will also drag and drop our modeled annual income
statements.
      In the course of time, the income statement presentations provided by the
companies you cover or monitor will change to incorporate one-time events or
permanent shifts. A familiar presentation might suddenly feature a line for
restructuring, impairment, or special charges related to a spin-off. As you
146   •   Ratio and Valuation Worksheet



incorporate this line item into your income statement P&L, remember to incor-
porate it into your annual income statement presentation on the R&V work-
book as well.
      For the balance sheet and cash flow statement, whether you link from sepa-
rate quarterly sheets or input directly on the R&V worksheet, align so that values
for each year line up with the annual accounts on the income statement. Even
more so than on the income statement, companies will insert new and one-time
or permanent line items on the balance sheet. If you are using separate quarterly
balance sheet and cash flow statement worksheets and the covered company adds
new line items that you input on those worksheets, be sure to adjust your R&V
presentation to accommodate those inputs.
      We will require annual balance sheet and cash flow data to complete the
modeled ratios for the two forward years. At this early stage in model building,
we recommend that you model the forward balance sheet and cash flows based
on their historical growth, at the appropriate multiple of GDP growth, or based
on the historical relationship between top-line growth and accounts in the bal-
ance sheet and cash flow statement.
      Once the model is complete, you can return to this process and refine your
financial statement modeling to integrate cash flow and balance sheet data and
to reflect more precise modeling of individual line items. Immediately, however,
you should incorporate new information that impacts the major balance sheet
and/or cash flow statement accounts. Figure 8.1 shows the annual financial state-
ment data from Celestica, a leading North American contract manufacturer.
When Celestica announced plans to exercise its option and use cash to retire
debt, we modeled the forecast changes in the appropriate accounts into our bal-
ance sheet. The revised cash and debt tallies linked from the R&V worksheet to
our income statement presentation will influence our net interest assumptions
for forward periods.


Ratios
Use of Ratios
At the very top of our R&V page, we compile a variety of ratios. Many will be used
directly in our calculation of dollar value of the asset. Some will be used to assess
the trajectory of the enterprise, across both its own history and in relation to
those of its peers. Some are gathered in one space primarily because they can then
most easily be added to the matrix valuation workbook for peer group compari-
son. All add value.
Figure 8.1

The annual financial statement stacks from the ratios and valuation of Celestica, a leading North America–based electronic manufacturing services (EMS) company.
The company’s actions and plans to reduce debt by paying it down from cash are reflected in the modeled balance sheet.


Celestica: Ratios, Valuations, Finances
Income Statement           1994       1998         2002        2003         2004         2005         2006         2007        2008        2009        2010
Revenues                    1,989.1    3,249.2      8,271.5     6,735.4      8,839.8      8,470.6       8,811.7     8,070.4     7,678.2     6,037.7     6,815.3
Cost of Goods               1,872.6     3,018.7     7,715.8     6,474.4      8,431.9      7,986.3      8,359.9      7,647.0      7,147.1    5,589.7     6,304.2
Gross Profit                   116.5       230.5       555.7        261.0        407.9       484.3         451.8        423.4       531.1      448.0        511.2
SG&A                           44.4       130.6       298.4       249.8         322.5       300.5         285.6        294.7      303.8       272.6       289.7
Amortization                       -       45.4         95.9        48.5         34.6         28.4          27.0        21.3        15.1        20.6        20.0
Acquisition/Accretion              -         8.1         3.0            -          3.5          7.9          0.9         2.3           -           -           -
Restructuring/Other                -       64.7        677.8       175.4        636.1        130.8        211.8         45.3        23.3        43.2        20.0
   Operating Income            72.1      (18.3)     (537.5)     (236.7)      (597.9)          16.7       (73.5)         59.8      188.9        111.6       181.5
Intrst, Other Expns              3.8       32.2         15.7          5.4         18.9        48.4         64.6         57.0        51.7        41.6        25.6
  P&L Prtx Income              68.3      (50.5)     (534.9)     (232.6)       (617.8)      (25.5)      (136.1)           8.6      146.4         70.3      155.9
Income Taxes                   26.3        (2.1)      (91.2)        33.2        245.2          21.3        14.5       (23.3)         5.0         3.7        28.1
Tax Rate                       39%          4%          17%        -14%         -40%         -84%         -11%       -271%           3%          5%         18%
   Net Income                  42.0      (48.5)     (443.7)     (265.8)      (845.4)       (46.8)      (150.6)          31.9       141.4        66.7      127.9
PF Net Income                                          216.5        (7.7)       104.2        140.6         93.5         61.9      188.5       114.9       168.0
Basic Shares Out               69.6       103.0        231.2       216.5        225.7        226.1        226.7        228.3      229.1       230.3       232.6
Diluted Shares Out              81.6      103.0       235.0         217.3       226.7        227.7        228.6        229.0      229.2       230.3       232.6         36.84
Rprtd Basic EPS                0.49      (0.47)      (1.92)      (1.23)        (3.76)       (0.21)      (0.66)          0.13      (3.15)        0.29        0.55
Rprtd Diluted EPS              0.49      (0.47)      (1.88)       (1.21)       (3.74)       (0.21)      (0.66)          0.13      (3.14)        0.29        0.55
 Adjusted Basic EPS            0.49        0.44         0.90      (0.10)         0.42         0.58         0.41         0.27        0.82        0.50        0.72
 Adjusted Diluted EPS          0.49        0.42         0.89      (0.10)         0.43         0.57         0.41         0.27        0.82        0.50        0.72
 EBIT                                    (18.3)     (537.5)     (236.7)      (597.9)          16.7       (73.5)         59.8      188.9        111.6      127.9
 EBITDA                                    68.7     (226.5)      (14.6)      (390.2)         169.4         60.7        190.6      298.1        211.6      228.9
 Enterprise Value                       1,072.7     4,135.4     2,136.0      3,226.7      2,462.8      2,220.8      1,182.6     1,023.3     1,552.5     1,558.3

                                                                                                                                                                   (continued)
Figure 8.1 (continued)


Celestica: Ratios, Valuations, Finances
Income Statement           1994       1998        2002        2003        2004        2005         2006        2007        2008        2009        2010
Cash                                               1,851.0     1,028.8       968.8       969.0        803.7      1,116.7    1,201.0      1,119.3      806.7
Balance Sheet              1994       1998        2002        2003        2004        2005         2006        2007        2008        2009        2010
Cash and ST Invstmnts                      31.7    1,851.0     1,028.8       968.8       969.0        803.7      1,116.7    1,201.0      1,119.3      806.7
Accounts Receivable                       463.0      785.9       771.5     1,023.3       982.6        973.2        941.2    1,074.0       808.9       825.1
Inventories                               430.9      775.6     1,030.6     1,062.9     1,058.4       1,197.9       791.9      787.4       634.3       647.0
Prepaid and Other Assets                   38.9      115.1       158.4        127.4       124.0        111.0      126.2         87.1        67.1        69.1
Incm Tx Rcvrbl                                                                 89.1       113.5         31.2        19.8        14.1        18.7        19.3
Deferred Income Taxes                      18.4        36.9        40.8         1.8         10.9         3.8         3.8         8.2         6.1         6.3
    Current Assets                        982.9    3,564.5     3,030.1     3,273.3     3,258.4      3,120.8     2,999.6      3,171.8    2,654.4     2,373.4
Capital Assets                            214.9      727.8       679.6       569.3       544.8         567.1      466.0       467.5       422.4       435.1
 Goodwill                                            948.0       948.0       872.9        874.5       854.8       850.5            -           -           -
Intangible Assets                         374.5       211.9       137.9      104.5          79.0        60.1        35.2        20.1        44.9        46.2
OtherAassets                               64.1      354.6       339.1       119.8         101.1        83.5       119.2      126.8       123.7       127.4
    Total Assets                        1,636.4    5,806.8     5,134.7     4,939.8     4,857.8      4,686.3     4,470.5     3,786.2     3,245.4     2,982.1
Debt                                                    6.9         3.4      503.4       750.9        750.8       758.5       733.1       583.3       255.6
Accounts Payable                          428.5      947.2      1,101.9     1,107.9     1,153.3      1,193.6    1,029.8     1,090.6       790.5       814.2
Accrued Liabilities                       174.9      475.4       382.3       486.6        492.1        487.9      402.6       463.1        312.5      321.9
Accnts/Pybl, Accrd                        603.3           -           -           -       119.9            -           -           -           -           -
Income Taxes Payable                       18.6        24.5         8.2        93.2          4.5        42.7        14.0        13.5        10.7        11.0
Deferred Income Taxes                       2.5        21.5        21.4         0.6          0.5         1.1           -           -           -           -
Cnvrtbl Debt                                                                                   -           -           -         0.2         0.2         0.2
Crrnt Prtn LT Debt                          2.3        2.7         2.7         2.6             -         0.6         0.2         1.0         0.1         0.1
   Current Liabilities                    626.7    1,471.3     1,516.5     1,690.9     1,770.3      1,725.9     1,446.6     1,568.4      1,114.0     1,147.4
Acrdd PstRtrmnt Bnfts                       6.3       77.2        86.0        81.0          76.8        54.9        70.4        63.2        66.9       68.9
Long-Term Debt                            133.5        4.2         0.7       500.8       750.9        750.2       758.3       732.1       583.2       255.5
Convertible Debt                                                             124.1             -           -           -           -           -           -
      Figure 8.1 (continued)


      Celestica: Ratios, Valuations, Finances
      Income Statement            1994      1998        2002        2003       2004        2005        2006        2007        2008        2009        2010
      Other Long-Term Liabilities                1.9         4.3         6.0       30.8        27.6        13.2        13.7         9.8         9.1           9.4
      Deferred Income Taxes                      8.7        46.2        57.2       23.4         17.8       47.5        63.3        47.2        35.6         36.7
         Total Liabilities                     777.2     1,603.2     1,666.4    2,451.0     2,643.4     2,591.7     2,352.3     2,420.7     1,808.8       1,517.8
      Net Cash                                           1,844.1     1,025.4     465.4        218.1        52.9       358.2       467.9       536.0        551.1
      Optn Cmpnt Convrtbl                                  804.6       603.5      210.2            -          -           -           -           -             -
      Debt
      Capital Stock                            912.1     3,670.6     3,297.8     3,559.1     3,562.3     3,576.6    3,585.2      3,588.5    3,588.7      3,696.4
      Warrants                                                                       8.9         8.4         8.4         3.1           -           -            -
      Cntrbtd Srpls                                           5.8      115.7       142.9       169.9       179.3       190.3       204.4       219.9       226.5
      Retained Incm/(Dfct)                    (52.2)     (294.7)     (578.8)   (1,473.6)   (1,545.6)   (1,696.2)   (1,716.3)   (2,436.8)   (2,412.3)   (2,284.4)
      Foreign Currency                          (0.6)        17.3       30.1        41.3        19.4        26.5        55.9         9.4        40.3         41.5
          Shareholders’ Equity                 859.3     4,203.6     3,468.3     2,488.8     2,214.4     2,094.6     2,118.2     1,365.5     1,436.6     1,679.9
         Liablts & Shrhldr Eqty              1,636.4     5,806.8     5,134.7     4,939.8     4,857.8     4,686.3    4,470.5      3,786.2    3,245.4       3,197.8

      Celestica
      Cash Flow Stmnt             1994      1998        2002        2003       2004        2005        2006        2007        2008        2009         2010
      Net Earnings                            (48.5)      -445.2      -265.8    (854.1)      (46.8)     (150.6)      (13.7)     (720.5)       66.7         127.9
      Dprctn & Amrtzn                           86.9       311.0       222.1      207.7       152.7       134.2       130.8       109.2      100.0         101.0
      Deferred Income Taxes                        -     (107.8)        27.1      234.6      (15.6)        55.2         6.4      (13.4)
      Accretion Cnvrtbl Debt                                                        17.6        7.6           -
      Nn-Csh Chg Optn Issnc                                     -        0.3         7.6        9.0         5.1         7.0          6.6
      Restructuring                                        194.5       (2.3)       35.3        11.0        47.9         5.1          1.1
      Other Charges                             64.7       292.1        80.5      482.4      (15.3)        34.6        14.0
      Settlmnt Cnvrtbl/Other                  (18.3)        (6.1)     (14.0)     (32.9)      (13.9)           -        18.0       850.3
      Invtry Writedown Rstrcng                                                     61.2                                            16.6
      Other                                                                          1.9       14.5          1.9
149




        Cash from Earnings                      84.9            -          -           -          -            -           -           -           -            -
                                                                                                                                                                    (continued)
Figure 8.1 (continued)


Celestica: Ratios, Valuations, Finances
Income Statement           1994       1998       2002        2003       2004        2005        2006       2007       2008        2009       2010
A/R                                                297.4         14.4    (253.0)       42.0       (24.8)       32.0    (132.8)
Invntrs                                            623.9      (252.6)       85.6                 (172.0)     406.0          4.5
Prpd, Othr                                          26.1       (43.2)     (12.9)         17.3        2.7      (6.8)        22.5
Incm Tx Rcvrble                                                           (50.0)      (24.4)        72.1       11.4         5.7
A/P & Accrd Lblts                                 (202.7)        65.2    (113.8)        51.2       108.0    (237.6)        58.9
Incm Tx Pybl                                         (0.4)        9.8       43.6        29.0      (75.1)     (21.2)       (0.5)
Noncash WC Changes                       (3.3)           -          -          -            -          -          -           -         -          -
   Cash frm Operations                    81.6      982.8     (158.5)    (139.2)       218.3        39.2      351.4      208.2      166.7      228.9

Acquisitions, Net Cash                  (48.7)    (111.0)       (0.5)     (39.6)        (6.5)     (19.1)
Purchase of Capital                     (65.8)    (151.4)     (175.9)    (142.2)     (158.5)     (189.1)     (63.7)     (88.8)    (120.0)    (120.0)
Asset Sale Prcds                             -       71.6         7.3      101.3         50.9        1.0       27.0        7.7
Other                                    (5.2)      (0.7)       (0.4)        0.6          2.2      (0.7)      (0.2)        0.3
   Cash Used Investing                 (119.7)    (191.5)     (169.5)     (79.9)      (111.9)    (207.9)     (36.9)     (80.8)    (120.0)    (120.0)

Bank Indebtedness                        (0.9)      (1.6)           -
Increase LT Debt                             -          -           -      500.0       250.0                  (1.4)
Repay LT Debt                          (423.2)    (146.5)       (3.5)      (41.1)       (3.4)          -      (0.6)          -           -          -
Debt Rdmptn Fees                                    (6.9)           -     (12.0)        (4.2)      (0.6)        3.5          -           -          -
Deferred Financing Costs                 (2.2)      (2.6)       (1.6)       (4.0)       (1.1)        5.3      (3.0)      (0.5)           -          -
Rprchs Nts/Cnvrtbl                                (100.3)     (223.5)    (299.7)     (352.0)       (1.3)          -     (30.4)           -          -
Issuance ST Debt                             -          -           -           -           -          -          -      (0.4)           -          -
Issuance of Share Capital                423.7        7.4         5.1        14.6         8.0          -          -        2.1           -          -
Repay ST Debt                                -          -           -           -           -          -          -          -           -          -
      Figure 8.1 (continued)


      Celestica: Ratios, Valuations, Finances
      Income Statement           1994       1998       2002      2003       2004      2005       2006       2007       2008       2009       2010
      Issue Shr Cptl                          (26.9)         -          -         -          -          -          -          -          -         -
      Rprchs Cptl Stock                                 (32.5)   (274.9)          -          -          -          -          -          -         -
      Dividends Paid                               -         -          -         -          -          -          -          -          -         -
      Other                                    (6.7)     (0.1)        4.2       1.3      (3.5)          -          -     (13.9)          -         -
         Cash Used Fncng                      (36.2)   (283.1)   (494.2)      159.1   (106.2)         3.4      (1.5)     (43.1)          -         -
                                                                        -         -          -          -          -          -          -         -
      Change in Cash                          (74.3)     508.2   (822.2)     (60.0)        0.2    (165.3)      313.0       84.3       46.7     108.9
      Cash, Beginning of Period                106.1   1,342.8    1,851.0   1,028.8     968.8       969.0     803.7     1,116.7    1,201.0   1,247.7
      Cash, End of Period                       31.7   1,851.0   1,028.8      968.8     969.0       803.7    1,116.7    1,201.0    1,247.7   1,356.5
      Cash & Short-Term Invstmt                    -
151
152   •   Ratio and Valuation Worksheet



      So far, we have three annual financial statements on our R&V page. This data
will contribute to our analysis but is not in and of itself our primary focus; we always
want to make best use of precious “eyeball” time. We will assume that the ratios por-
tion will occupy approximately 40 to 45 lines at the top of our R&V page.
      We will include up to 30 highly useful ratios in this section. We lack space
to give each its full due, but for each ratio we provide a cursory description of its
value and use. For a highly useful, compact, and yet thorough analysis of ratios,
please see The Analysis and Use of Financial Statements by White/Sondhi/Fried.
      Ratios can be grouped thematically; we do so as much for consistency as for
convenience. In nearly every case as we go down this list, we will construct our
ratios from the data available in the annual financial statements that are either
input or linked to the R&V page.
      In general, the income statement and the cash flow statement measure per-
formance across a period of time. The balance sheet, by contrast, provides data
on the status of the enterprise at a single point in time. In our ratios, we want to
make the static data from the balance sheet approximate performance across a
period, thus better aligning with data coming from the income statement and
cash flow statement. Accordingly, when we use balance sheet data in our ratios,
we will generally use an average of two periods, to best mimic the period-span-
ning nature of the income statement and cash flow statement data.
      Keep in mind that financial definitions are not hard and fast; various inves-
tors may use the same terms to mean different things. As much as possible, we
hew to the widest conventional use. In addition to describing and explaining how
to construct the ratio, we’ll try to lend some real-world color when applicable.
      From a practical perspective, we will construct our ratios in a single “strip”
down the page. Once we’ve completed the process, we will drag and drop for all
periods.


Internal Liquidity
The internal liquidity ratios provide information on the immediate cash avail-
ability or liquidity of the enterprise and also, in the measurement across multiple
periods, how consistently a company maintains its liquidity. The half dozen we
most closely track follow.


Working Capital
Working capital (WC) measures the difference between a company’s current
assets and current liabilities. Historically, high levels of working capital (e.g.,
              Ratio Analysis, Part 1: Internal Liquidity and Operating Efficiency   •   153



excess of current assets over current liabilities) have been regarded positively
because they show a company as being highly liquid to meet immediate cash
needs. In a downturn, high working capital can serve as a kind of “bank” for the
well-managed enterprise; key current asset accounts (principally, accounts receiv-
able and inventories) can be drawn down, which is contributory to cash flow and
may help keep cash flow positive amid negative profit performance. Poor work-
ing capital management in a down cycle, on the other hand, is seen by investors
as an indictment of management.
      To calculate working capital in a cell, let’s assume that for 2007 current
assets are in cell M125 and current liabilities are in cell M140. The working capi-
tal measure for that snapshot in time (most typically December 31, 2007) would
be M140 M125. To simulate the period-spanning nature of an ongoing enter-
prise, we use an average drawn from the current end of period and the prior end
of period. There are a few ways to do this, such as using AVERAGE or (n n1)/2,
where n equals the prior period. For consistency, I mainly use AVERAGE. To cal-
culate working capital, the formula in this instance would be AVERAGE(L140
:M140) AVERAGE(L125:M125).


Trade Working Capital
You won’t find the term “trade working capital” cited in discussions of ratio
analysis. Working capital plays a key role in discounted free cash flow calcula-
tion—but not working capital per se. Instead, the period-over-period change in
working capital is contributory (or depletive) to free cash flow. But if a company
has generated cash in its current account, should that company be penalized by
this seeming “use” of cash in the DFCF calculation? For many analysts, the
answer is no. From a practical perspective, you will find that many analysts adjust
working capital so that companies are not penalized in cash flow valuation for
good practices such as cash generation. CFA Institute, for example, in its mea-
surement of free cash flows, uses the change in working capital excluding cash
and short-term debt.
       Current assets and liabilities contain various accounts that are unpredict-
able and variable in nature; these include deferred tax assets, income taxes receiv-
able, income taxes payable, and deferred revenues. Some of these accounts are
influenced by financial events rather than operating events. In our calculation of
working capital for use in DFCF, we take it a step further and include only the
principal operating line items in current accounts. These are accounts receivable
and inventories from current accounts, and accounts payable from current
liabilities.
154   •   Ratio and Valuation Worksheet



     To calculate trade working capital in a cell, sum the average of two-period
accounts receivable and inventories and subtract the average of two-period
accounts payable.


Current Ratio
Current ratio shows the difference between current accounts and current liabili-
ties, expressed as a ratio. Historically, for most sectors, investors want to see posi-
tive current ratios; they would prefer to see current accounts at two or more times
the level of current liabilities (a current ratio of 2.0). When the economy tanks,
however, investors want to see current ratio decline as companies improve col-
lections and cut inventories to free up and preserve cash. There are, of course,
some variations in these trends based on industry specifics. Sectors with nontypi-
cal current ratios can include finance and energy; industries can include retail
and other consumer-sensitive areas.
       To calculate current ratio in a cell, divide the multiperiod average of current
liabilities by the multiperiod average of current assets. Current ratio is an exam-
ple of a ratio that will not figure specifically in the calculation of dollar value
of the asset. But it has a lot of informational value. Current-ratio changes across
the cycle speak volumes about management responsiveness in a dynamic
environment.


Quick Ratio
Quick ratio is considered by some investors a better gauge of liquidity and cash
availability than current ratio. It sums real and perceived liquidity and measures
this figure against current liabilities. Quick ratio has answered to multiple defini-
tions over the years. Most typically, it is defined as the sum of cash and accounts
receivable divided by current liabilities. Some investors and analysts (e.g.,
Investools.com) calculate quick ratio as current assets less inventories divided by
current liabilities. However it is measured, creditors like to see a quick ratio
exceeding 1; again, there can be some sector-based and industry-based variability
in average ratios.
       To calculate quick ratio in a cell, we use the traditional methodology. Divide
the multiperiod average of cash plus accounts receivable by the multiperiod aver-
age of current liabilities. Quick ratio is an example of a ratio that we do not use
in the calculation of dollar value of the asset. It has informational value, particu-
larly in times of financial distress. It is also useful when viewed in comparison to
the peer group within the matrix workbook.
              Ratio Analysis, Part 1: Internal Liquidity and Operating Efficiency   •   155



Working Capital/Sales Ratio
The working capital/sales ratio is not as intuitive as the other ratios; it is not
directly applicable to DFCF calculations; it also does not figure in our determina-
tion of dollar value of an asset. Still, it has useful information content. Analysts
and investors track this figure to gauge a company’s ability to finance a higher
level of sales from existing capacity—that is, without incurring additional short-
term debt or (CP commercial paper is a form of very short-term financing used
to meet day-to-day expenses). Watching this ratio is particularly useful in times
of economic distress. As the economy weakens and company revenues decline in
tandem, working capital should be reduced, to lessen strains on cash use and to
help maintain positive cash flows. Thus, investors want to see this ratio hold
constant in trying times. A decline in the ratio would signal distress at the com-
pany itself and an inability to fund sales at current levels.
      To calculate the ratio of working capital to sales in a cell, divide working
capital (calculated a few cells above) by sales for the given period. Figure 8.2
shows the internal liquidity ratios for Texas Instruments, a cash-rich company
with a solid quick ratio.


Operating Efficiency
The operating efficiency ratios measure execution and efficacy of asset use. The
key inputs measure turnover time, including days sales outstanding, and provide
information on the cash cycle. A cash-rich company operating in high-margin
niches can mask a flawed operating strategy or sloppy execution; the operating
efficiency ratios provide early warnings on such companies.


Receivables Turnover
Receivables turnover describes the duration of accounts receivable, or how long
receivables are held in relation to revenue. A high turnover figure is desirable,
because it shows that the enterprise collects its receivables efficiently and effec-
tively and that it does business with good credits. Low receivables turnover ratios
may signal struggling customers and can be a prelude to the use of factoring
companies to securitize or otherwise take poor-quality receivables off the balance
sheet at a discount. Receivables turnover is also viewed as a good indicator for
cash flow and for operating performance in general.
      In the internal liquidity ratios cited so far, which have all been based on
balance sheet data, we find it useful and informational to take a multiperiod
Figure 8.2

Dallas-based Texas Instruments had $2.6 billion in cash and no debt as of mid-2009. Accordingly, its liquidity measures were top of the industry.

Texas Instruments
RATIOS                                   1998          1999                        2005         2006          2007          2008         2009E         2010E
  Internal Liquidity
Working Capital                              2,776         3,656                      6,839         5,776         4,893         4,258         4,272       4,267
Trd Wrkg Captl                               1,468          2,116                     2,335         2,651         2,503         1,964         1,886       1,952
Current Ratio                                   2.2           2.4                        3.9           3.8           3.4           3.8           3.9         3.8
Quick Ratio                                     0.9           1.0                        1.3           1.4           1.5           1.3           2.0         2.0
Working Capital/Sales                           0.3           0.4                        0.5           0.4           0.4           0.3           0.4         0.3
               Ratio Analysis, Part 1: Internal Liquidity and Operating Efficiency   •   157



perspective. For those ratios that mix balance statement line items such as receiv-
ables with income statement data such as revenue, it is essential to take a multi-
period approach with balance statement line items.
      To calculate receivables turnover in a cell, divide revenue for a given year by
the two-period average of accounts receivable, using most recent annual period
and nearest prior annual period.


Receivables Collection Period
The information value of receivables turnover is enhanced when we calculate the
receivables collection period, also called receivables days outstanding. This tells
us how many days are required to turn over average receivables. Information
measured in days is essential to determining days sales outstanding and cash
cycle, widely viewed as measures of operating efficiency and—indirectly—the
quality and solvency of customers.
      To calculate receivables collection period in a cell, we use the value calcu-
lated immediately above—receivables turnover—and divide it by the days in the
period, or 365 days. There are variations on this measurement period. Some
investors and analysts round to 360 days; others may use the number of business
days in a year (about 250). If you use anything other than 365 days, the point is
to be consistent throughout all your models. However, use of anything besides
365 days erodes the comparable value of your measures when you bring them
outside the self-contained world of your models.


Total Asset Turnover
In between our analysis of receivables turnover and inventory turnover, we mea-
sure total asset turnover. This is just what it sounds like: total revenues divided
by total assets. This broad measure stands alone and does not contribute to sales
cycle information such as days sales outstanding or cash cycle. But it is an infor-
mation-rich reading on the underlying efficiency of an enterprise.
      Some investors and analysts are more interested in a cleaner asset return
number, so they strip out intangibles and goodwill from total assets. As always,
if you are going to make this adjustment, be sure to do so consistently throughout
your coverage universe; and make sure any comparisons made outside the uni-
verse are against like-adjusted measures.
      To calculate total asset turnover in a cell, divide revenue for a given year by
the two-period average of total assets, using most recent annual period and near-
est prior annual period.
158   •   Ratio and Valuation Worksheet



Inventory Turnover
Much as receivables turnover provides insight on the quality of the customer base
as well as a company’s efficiency in collecting bills, inventory turnover is more
than just a number; it provides insight on how well the supply chain is function-
ing and how optimally procurement officers are making purchases. The impor-
tant distinction between accounts receivable turnover and inventory turnover is
in the denominator: inventories are measured as a multiple of cost of goods sold,
not of revenues.
      To calculate inventory turnover in a cell, divide cost of goods sold for a
given year by the two-period average of total inventories, using most recent
annual period and nearest prior annual period.


Inventory Collection Period
Again mimicking receivables turnover, the information value of inventory turn-
over is enhanced when we calculate the turnover collection period (also called
inventory days outstanding). Inventory collection period is also a referendum on
the quality and solvency of customers, pricing power, market share trends, and
many other useful (though not easily quantifiable) data points.
      To calculate inventory collection period in a cell, we use the value calculated
immediately above—receivables turnover—and divide it by the days in the
period, or 365 days. There are variations on this measurement period. Some
investors and analysts round to 360 days; others may use the number of business
days in a year (about 250). If you use anything other than 365 days, the point is
to be consistent throughout all your models. However, use of anything besides
365 days erodes the comparable value of your measures when you bring them
outside the self-contained world of your models.


Days Sales Outstanding
Cash cycle is simply another way of saying receivables turnover period or receiv-
ables days outstanding. An alternate way to record days sales outstanding (DSO)
in a single cell is to divide accounts receivable for the period by revenues and
multiply the result by the days in the period. In our ratios presentation, we have
been careful to maintain ratios based on a full year. Sell-side analysts and inves-
tors who model more meticulously may want this DSO information on a more
timely or quarterly basis. For them, we recommend creating space within the
              Ratio Analysis, Part 1: Internal Liquidity and Operating Efficiency   •   159



quarterly balance sheet workbook to perform this calculation. You’ll need to
import quarterly revenue data from the income statement presentation and then
proceed as below.
     To calculate days sales outstanding in a cell, divide accounts receivable for
the period by revenue for the period and multiply the result by days in the period.
For quarterly measurements, analysts typically use 90 days. For annual measure-
ments, analysts may use 360 or 365 days. Whatever the measurement period,
remember to use it consistently across all your covered companies.


Fixed Asset Turnover
The next two ratios—fixed asset turnover and equity turnover—are mainly for
informational and comparison purposes; they do not figure in the calculation of
dollar value of the asset. They are mainly tracked on a company-specific basis for
variance from the historical norm and within the industry comparison matrix
for changes in the historical relationship to the peer group.
      Fixed assets are tangible assets, mainly property, plant, and equipment
(PP&E). Fixed asset turnover indicates the revenue generated by the money spent
on PP&E. Because PP&E is depreciated over time while sales presumably are
growing, in a normally progressing company this ratio should be increasing. If
this ratio flattens out or begins a sustained fade (as opposed to a cyclical blip),
that can be a sign that a company has invested unwisely in unproductive assets.
      To calculate fixed asset turnover in a cell, divide net sales for the period by
the average (two-period) PP&E taken from the balance sheet.


Equity Turnover
Shareholders’ equity is share capital (money invested in the company at incep-
tion, such as par value of common stock and additional paid-in capital) plus
retained earnings (net income after dividends) minus treasury shares (costs to
repurchase stock); a few other items such as accumulated other comprehensive
income and foreign exchange can impact this number. Alternately, shareholders’
equity can be calculated by what it isn’t: assets less liabilities. As mentioned, we
contend that FAS 142, the accounting board’s decision to impair rather than
depreciate goodwill, has had consequences for shareholders’ equity that far out-
weigh the rationale for the standard’s imposition. These include reducing the
informational value and even validity of stockholders’ equity in ratios such as
debt to capitalization and return on equity.
160   •   Ratio and Valuation Worksheet



      Despite these distortions to shareholders’ equity, the fact that they’ve mani-
fested across the entire universe of public stocks creates a common ground for
comparison. To calculate equity turnover in a cell, divide net sales for the period
by average (two-period) stockholders’ equity taken from the balance sheet.


Payables Turnover
Payables turnover and payables days outstanding (PDO) are our final operating
efficiency ratios. They play a role in determining cash cycle, perhaps the single
best measure of operating efficiency. But payables turnover and PDO have infor-
mational value all their own. In a normally functioning economy, a company’s
ratio of what it pays out to its vendors should about track the rate of sales growth.
In a weakening economy, payables will decline; in a recovering economy, payables
will rise as companies prep for higher activity levels. Because revenue recognition
can introduce lags, cyclical changes can distort payables turnover. We are there-
fore more interested in the long-term trend. A deterioration in payables turnover
would signal acquired consumables (i.e., payments to vendors) are not generating
the necessary level of sales growth.
      To calculate payables turnover in a cell, divide net sales for the period by
average (two-period) accounts payable taken from the balance sheet.


Payables Collection Period
Again mimicking receivables turnover, the information value of payables turn-
over is enhanced when we calculate the turnover collection period (also called
payables days outstanding). This is the final step in advance of calculating cash
cycle, viewed as one the best measures of operating efficiency. It is also a referen-
dum on the quality and solvency of customers, pricing power, market share
trends, and so on.
      To calculate payables collection period in a cell, we use the value calculated
immediately above—payables turnover—and divide it by the days in the period,
or 365 days. There are variations on this measurement period. Some investors
and analysts round to 360 days; others may use the number of business days in a
year (about 250). If you use anything other than 365 days, be consistent through-
out all your models. However, use of anything besides 365 days erodes the com-
parable value of your measures when you bring them outside the self-contained
world of your models.
              Ratio Analysis, Part 1: Internal Liquidity and Operating Efficiency   •   161



Cash Cycle
Cash cycle, sometimes called cash conversion cycle, measures the length of time
taken to convert inputs into cash flows. More prosaically, it measure the time
between the purchase of raw materials and the collection of accounts receivable
on items or services sold. Cash cycle aggregates the efficiency of accounts receiv-
able, inventories, and accounts payable. Accordingly, it draws on the days out-
standing calculations we have derived above for all three of these accounts.
      Intuitively, we would like to see cash cycle shorten rather than lengthen. We
have already discussed how the economic cycle can affect these measures. In
tough times, companies typically seek to mitigate downshifts in operating cash
flow by reducing working capital. Given lags in revenue recognition, this can have
short-term distortive effects on cash cycle. So again, we are less concerned with
cyclical blips and more concerned with structural lengthening or shortening in
the cash cycle.
      One effect we have not discussed relates to regional patterns. In Europe,
receivables and payable terms tend to be longer than in the United States; terms
in Asia as of this writing are not so easy to categorize. Firms in most sectors
(exceptions can be financial and utility) have global ambitions. Accordingly, the
careful analyst or investor needs to assess changes in cash cycle within the con-
text of changes in regional business mix.
      To calculate cash conversion cycle in a cell, add receivables days outstanding
and inventory days outstanding and subtract payables days outstanding.
      The operating efficiency of Singapore- and California-based Flextronics
International is on display in Figure 8.3. Flextronics operates massive global
manufacturing campuses and produces high-velocity items such as mobile phones
and PCs. As such, the company is meticulous in driving down its cash cycle,
which now runs at one third to one quarter of peak levels.
      We’ve now determined key elements of the company’s liquidity and operat-
ing efficiency. Next we’ll look at what ratio analysis tells us about profitability
(return), as well as cash-generating capacity.
Figure 8.3

Operating in the EMS industry, where razor-thin margins prevail, Flextronics has been relentless in driving down its cash cycle.

Flextronics International: Ratios, Valuations, Finances
RATIOS                         1994     2000       2001        2002      2003      2004       2005      2006      2007       2008      2009     2010      2011
  Operating Efficiency
Receivables Turnover                        6.66      7.33        7.02      9.44       7.76      8.60     10.36      10.74      7.76    13.36    12.52     14.58
Receivables Collect Period                    55        50          52        39         47        42        35         34        47       27        29       25
Total Asset Turnover                        1.32      1.60        1.52      1.59       1.52      1.44      1.42       1.53      1.41     2.73     2.49      2.90
Inventory Turnover                          5.27      6.23        9.46     11.08      11.62      9.73      8.33       6.96      6.31     9.85     9.20     10.72
Inventory Turn Time (days)                    69        59          39        33         31        37        44         52        58       37       40        34
Days Sales Outstanding                       124       108          91        72         78        80        79         86      105        64       69        59
Fixed Asset Turnover                        5.24      6.62        6.45      6.81       8.94      9.29      9.77       9.43     11.18    13.26    11.62     13.53
Payables Turnover                           5.56      8.18        6.68      8.35       6.77      6.28      5.62       5.48      5.19     7.64      7.12     8.30
Payables Turn Time (days)                     66        45          55        44         54        58        65         67        70       48        51       44
Equity Turnover                             2.59      3.00        2.94      2.95       3.32      3.03      2.89       3.05     3.38     16.87    14.82     15.34
Cash Cycle                                    58        64          36        28         25        22        14         20        35       17        18       15
                                                  Chapter 9
           RATIO ANALYSIS,
           PART 2:
           RETURN RATIOS AND
           CASH FLOW RATIOS




The return ratios on our R&V page include some tallies we’ve calculated else-
where, mainly measurements of margin, as well as some return ratios not yet
measured such as return on assets (ROA) and return on equity (ROE). The previ-
ously calculated margin ratios provide value on the R&V page by being available
in a compact format. This makes it visually easier to track expansion or contrac-
tion in the margins. Equally important, this array best lends itself to importation
into the industry matrix workbook, where we can track margin trends not just
for the individual equity but in relation to the historical relationship to the
group.


Margin Ratios
While these margin-based ratios exist elsewhere in our workbook, we won’t link
to the existing calculations to display these ratios. Instead we’ll derive them as
formulas directly in the cell. There are two reasons for doing so. One, while links
are indispensible to our calculation of dollar value of an asset, every link is
another place where information can go awry. Links have been known to imper-
fectly adjust if the linked-to or linked-from workbook is not open when changes

                                                                                163
164   •   Ratio and Valuation Worksheet



are made on its counterpart. We’ll discuss this issue in more detail in Part 4, our
discussion of relational valuation. At that point we’ll be we transferring data
between individual company workbooks and an industry “matrix” workbook—
and the risk of links-gone-awry will intensify. Two, links do not always lend
themselves to drag-and-drop replication, particularly if the linked-to columns
are consecutive while the linked-from columns are dispersed.
      We’ve discussed margins extensively in Part 1 on income statement presen-
tation, so our discussion here will be cursory. But we do want to hit on a few key
points and tendencies.


Gross Margin
Gross margin measures the percentage of gross profit left after cost of goods is
deducted from revenues. Generally, gross margin will be lower for companies
with capital intensity and high fixed costs; lower gross margins are associated
with companies that do their own manufacturing. Higher gross margins are
associated with lower capital intensity, outsourced production, and a higher pro-
portion of variable costs; information processing and business services are exam-
ples of industries with higher gross margins. Accordingly, analysis of gross
margin trends across the company or peer group supersede the value of analyzing
trends in relation to all sector and industries. Because gross margins play such an
enormous role in operating margin, pretax margin, and net margin, the same
basis of comparison holds true for all margin measures.
      All the inputs needed for margin calculation are in the annual income state-
ment data imported onto the R&V page. Because income statement data expresses
performance across a span (as opposed to the snapshots provide by the balance
sheet), it is sufficient to use one period only rather than the average of two
periods.
      To calculate gross margin in a cell, divide gross profit for the period by
revenue for the period (in both cases, for the full year), and express it as a
percentage.


Operating Margin
Operating margin ideally indicates a company’s profitability purely from the
course of its business and to the exclusion of inputs related to prior financial
decisions (i.e., interest paid on debt or earned on cash) or nonoperating events
(i.e., taxes along with distortions in tax rates related to impairments, restructur-
ings, or other events).
                     Ratio Analysis, Part 2: Return Ratios and Cash Flow Ratios   •   165



      Depending on the nature of your coverage universe, you may want to create
space for pro forma operating margin. Remember, however, that the value of
margin analysis is not in isolation but in relation to historical trends within the
individual company and in relation to the peer group. For this reason, even
though my most recent coverage has been in technology—an industry where
adjusted earnings and valuations are the rule rather than the exception—for this
section of the R&V page I pretty much use GAAP margins. Also be aware that if
you enter margin data into your matrix workbook, it must have measurement
consistency. Because all firms provide GAAP data but not all provide pro forma
data, GAAP values should be used at any place where trends within the matrix
supersede the value of stand-alone analysis.
      To calculate operating margin in a cell, divide operating income for the
period by revenue for the period, and express it as a percentage.


Pretax Margin
Pretax margin is effectively operating margin adjusted for nonoperating financial
structure strategy and decisions. The primary adjustment levers are interest
income and interest cost, but they are not the only factors. Line items between
operating income and pretax income can include effects of other financing deci-
sions, including debt prepayments and net realized investment gains, as well as
foreign exchange effects, and sundry/other.
      To calculate pretax margin in a cell, divide pretax income for the period by
revenue for the period, and express it as a percentage.


Net Margin
Net margin represents net income as a percentage of revenue. Line items that
contribute to the difference between pretax income and net income of course
include taxes, but also may include equity income, minority interest, earnings (or
loss) from discontinued items, accounting changes, and other.
      To calculate pretax margin in a cell, divide pretax income for the period by
revenue for the period, and express it as a percentage.


EBITDA Margin
EBITDA margin measures earnings before interest, taxes, depreciation, and
amortization (EBITDA) as a percentage of revenue. EBITDA and EBITDA margin
are seen as the best measure of the pure cash-based operating performance of the
166   •   Ratio and Valuation Worksheet



company, as they exclude the two most significant noncash items (depreciation
and amortization) from the return calculation.
      Personally I’ve always been wary around EBITDA margins. I can remember
attending a company presentation in which the CEO was touting growth in
EBITDA—while failing to mention that the $800 million in interest costs incurred
in the acquisition binge that drove EBITDA growth was larger than operating
income in any prior year. (The CEO and his firm were put out of their misery
when the assets of the collapsed company were subsequently sold at fire-sale
prices.)
      Nonetheless, EBITDA does figure in two prominent valuation equations.
The first is in determining values for mergers and acquisitions (M&A) pur-
poses. In analyzing acquisition of an asset, the key concern is with cash-based
operating characteristics; nonoperating items, such as interest cost and taxes,
have little role in the analysis. In the normal course of business, large firms
acquire small ones. Interest costs that seem prohibitive for a small firm can
vanish into the folds of a much larger firm; similarly, the tax structure of the
larger firm will supersede that of the smaller firm. Accordingly, analysts tend
to evaluate proposed acquisitions on the basis of multiples of revenue and
EBITDA.
      Before calculating EBITDA margin in a cell, you need first to compile
EBITDA. Every income statement presentation is different or at least has some
different elements. There is no one-size-fits-all formula that we can use in our
ratio section. So we find it useful to compile EBITDA for each year immediately
under the appropriate annual income statement presentation. To calculate
EBITDA margin in a cell, use that EBITDA value, divide it by revenue for the
period, and express it as a percentage.


EBITDA/Enterprise Value (EV)
The second highly efficacious use of EBITDA is within the EBITDA/enterprise
value (EV) calculation. EBITDA/EV is in that rare valuation ratio seen as cutting
across all sectors and industry classes. In terms of stand-alone valuation meth-
odologies, it is likely second only to P/Es, and it is preferred over P/Es by profes-
sional investors. Like EBITDA, we first calculate and situate enterprise value for
each year under the appropriate annual heading immediately below the income
statement.
     Enterprise value at its simplest is the market value of equity (share price
times shares outstanding) plus the market value of debt minus any cash and
                      Ratio Analysis, Part 2: Return Ratios and Cash Flow Ratios   •   167



equivalent holdings. EV is felt to give the fullest picture of valuation because it
incorporates input from all holders (including stock and bond investors).
     To calculate EBITDA/EV in cell, divide EBITDA by EV, and express as a
percentage.


Enterprise Value (EV)/EBITDA
The inverse of EBITDA/EV (which is expressed as a percentage), EV to EBITDA
is expressed as a floating-point number. Much as high EBITDA/EV percentage
values are seen as superior to low values, low EV/EBITDA values are superior to
high values.
      To calculate EV/EBITDA in a cell, divide EV by EBITDA, and express it as
a number with at least one digit to the right of the decimal.


Return Ratios
The return ratios we calculate figure both in our assessment of the ongoing enterprise
and in our calculation of the dollar value of the asset. Again, we are mainly monitor-
ing these to determine the ability to widen return ratios in a normally functioning
economy, or at least maintain them at appropriate levels. One of the ratios, return on
equity, figures prominently in discounted free cash flow analysis.


Return on Invested Capital (ROIC)
Investors in the aggregate are likely to overemphasize P/E ratios in their invest-
ment decisions, because P/Es are easily understood and because earnings are
widely modeled; that enables P/E ratios to encompass new information easily and
quickly. Simultaneously, many investors distrust the quality of information con-
tained in accounting earnings and the imperfect and sometimes speculative
information contained in the stock price. These investors have turned to more
cash flow–based calculations of value.
      One method for measuring historical cash flow performance of a company
is return on invested capital (ROIC). The value of ROIC is that it can be used to
provide a consistent measure of cash-based return of a company over an extended
period. Single-period cash flows can be manipulated by management, but cash
flow over an economic cycle is more difficult to massage.
      ROIC shares the common problem of any ratio with more than two inputs:
variability in calculation. Some sources define ROIC as net income minus divi-
168   •   Ratio and Valuation Worksheet



dends divided by total capital. Many firms are not dividend paying; anyway, this
method does not satisfy the cash-conscious investor.
      On the other hand, ROIC is used in the Stern-Stewart Calculation of Eco-
nomic Value Added (EVA), and this ROIC has some variations of its own. EVA
advocates believe that companies whose ROIC exceeds their weighted average
cost of capital are creating value.
      Our preferred method of calculating return on invested capital is net oper-
ating profit after taxes, or NOPAT, divided by adjusted assets. NOPAT, the
numerator in this equation, is net operating earnings before interest and amor-
tization minus cash taxes. For the denominator, we adjust total assets by sub-
tracting cash and subtracting non-interest-bearing current liabilities—that is, all
current liabilities except current debt-related items (e.g., notes payable, short-
term debt, and debt due within one year) as well as accrued liabilities. Cash taxes
can be found at the bottom of most cash flow statements as supplementary
information.
      Because of this reliance on cash taxes, we need to adjust our cash flow state-
ment model to encompass this line item. Gather historical cash taxes, and list
them immediately below your formal cash flow presentation; we also include a
line to represent cash taxes as a percentage of accounting or book taxes. Because
accounting earnings use straight-line depreciation while cash taxes are based on
accelerated depreciation, you can assume that for a normally functioning (i.e.,
nonimpairing) company, accounting taxes will usually exceed cash taxes.
      To calculate ROIC in a cell, our numerator consists of net income with
added-back interest and amortization (the former brought up from the annual
income statement, the latter from the cash flow statement), minus cash taxes
(listed immediately below the formal cash flow statement). The denominator
consists of total assets minus cash and equivalents (from current assets on the
balance sheet) and non-interest-bearing current liabilities as described
above.
      ROIC is seen as a measure of historical performance of a company, and
purists might gasp at a forward calculation of this ratio. But we are already mod-
eling nearly every input needed, and cash taxes can be expressed as a percentage
of book taxes. In my experience, cash taxes tend to run at 55 to 70 percent of book
taxes for a normally functioning company. Using these inputs, we can model a
company’s ROIC one or two years in advance. While this has less informational
value than measuring historical ROIC, we can project whether a company will
continue to exceed (or lag) its weighted average cost of capital and thus create (or
destroy) economic value.
                     Ratio Analysis, Part 2: Return Ratios and Cash Flow Ratios   •   169



Return on Total Capital
For the generalist, this ratio is simpler to calculate and, with fewer variables,
inputs, and adjustments, may have more forward-looking and predictive value
than estimated ROIC. Return on total capital is akin to EBITDA/EV, but it uses
a more straightforward calculation. Even so, adjustments are needed for a true
definition.
      One definition of return on total capital is net income divided by the book
value of equity and the book value of debt. Alternatively, some investors use
NOPAT or EBITDA in the numerator, and some may use the market value of debt
or the market value of equity. Again, we stress that consistency across the uni-
verse, and when comparisons are made outside the universe, the methodology
must conform to the comparison universe.
      We use a very straightforward ratio: net income divided by the sum of
equity and debt, both at book value. To calculate return on total capital in a cell,
bring net income for the year up from the income statement and divide by the
sum of the book value of total debt and shareholders’ equity.


Return on Assets
Return on assets (ROA) is a very well known financial ratio and perhaps the
second-most-common in valuation analysis, after return on equity. It is a straight-
forward ratio built from well-known inputs in the financial statements: net
income divided by average (two-period) assets.
     The chief danger in misusing ROA is its lack of applicability across industry
groups. Financial firms and insurers in particular tend to have low ROA; technol-
ogy firms and software firms in particular—really, any industry in which low
capital intensity is the norm—tend to have high ROA. So, more so than with
almost any ratio, ROA has the most value when used to compare return perfor-
mance across carefully constructed universes of like entities.
     To calculate return on assets in a cell, divide net income for the period by
the two-period average of total assets.


Return on Equity: Two Methods
Later, we’ll examine in some detail what we regard as dangerous patterns in
stockholders’ equity related to company practice (i.e., risk-taking) as well as
bureaucratic blunder (i.e., impairment rather than depreciation of goodwill). For
170   •   Ratio and Valuation Worksheet



now, and without passing judgment, we will focus on calculating return on equity
(ROE), given its paramount place in the valuation pantheon. ROE, to paraphrase
(in the loosest sense of the word) Winston Churchill, is the worst of the valuation
methodologies, except for all the others. Its strengths, like those of the P/E, are
its universal applicability, its familiarity to professional and casual investors, its
ease of calculation and interpretation, and its widespread use that cuts across
sectors and nations.


Standard ROE
The most straightforward method of calculating ROE is to divide net income
by stockholders’ equity, or conversely net income per share by book value (i.e.,
stockholders’ equity per share). Because share count and thus EPS can be
affected by financial strategy decisions—that is, nonoperating inputs—for
comparison over extended periods it is much more useful to divide net income
by stockholders’ equity; that is what we do in the ratio section on our R&V
page. To calculate ROE in a cell, divide net income by the two-period average
of stockholders’ equity.


DuPont ROE
Professional investors also calculate return on equity using the DuPont method.
We do this on our stock value worksheet, which we’ll discuss later on. There are
variations on the DuPont method; we use a three-stage DuPont. The advantage
of this three-stage method is that it gives a visible presentation of the three stages
of the ratio calculation, providing additional input.
      To calculate DuPont ROE in a cell, in three separate cells divide (1) net
income by sales, (2) sales by assets, and (3) assets by equity. This will produce
three percentage calculations; in a fourth cell, multiply these three percentages
together. Again, for inputs coming from the balance sheet (i.e., assets and stock-
holders’ equity), use two-period averages. For a cross-check, compare this out-
come to the ratio produced by net income divided by stockholders’ equity; if they
vary, then one of your inputs is out of whack.
      Figure 9.1 shows return measures for Analog Devices Inc. Targeted benefits
from the company’s restructuring efforts and asset dispositions in low-return
niches have been superseded by the economic tsunami. In the absence of major
goodwill impairment, Analog Devices Inc. will have to rebuild ROE the old-
fashioned way, by earning it.
Figure 9.1

Analog Devices Inc. participates in the high-margin semiconductor space, specializing in analog chips. After a dip in the October 2009 year, the company is
looking to rebuild its margins going forward.

Analog Devices
RATIOS                         1997       1998        2002       2003        2004       2005        2006       2007        2008       2009E      2010E
   Return Ratios
Gross Margin                                47.8%       53.0%      54.9%       59.0%      57.9%       58.5%      61.2%       61.1%      55.0%       57.8%
Operating Margin                            12.8%        6.9%      18.3%       26.9%      21.6%       21.4%      23.1%       24.2%      12.1%       18.2%
Pretax Margin                               12.2%        8.2%      18.7%       28.2%      24.6%       25.7%      26.8%       25.8%      12.7%       18.0%
Net Profit Margin                             9.7%        6.2%      14.6%       22.0%       17.4%      21.3%      20.5%       35.4%      10.4%       13.9%
Return on Invested Capital
Return on Total Capital                     21.2%        3.7%       9.6%       16.4%       11.1%      15.5%      17.5%       38.5%       8.4%       11.2%
Return on Assets                            12.8%        2.1%       6.6%       13.2%        9.1%      13.1%      14.5%       30.2%       6.5%        9.1%
Return on Equity                            21.2%        3.7%       9.6%       16.4%       11.1%      15.5%      17.5%       38.5%       8.4%       11.2%
EBITDA Margin                               22.6%       22.2%      26.9%       34.0%       31.1%      32.4%      33.1%       31.5%      20.4%       25.2%
EBITDA/Enterprise Value                      6.0%        3.2%       4.5%        5.1%        5.3%       6.7%       7.0%        9.9%       5.0%        7.0%
Enterprise Value/EBITDA                       16.5        31.2       22.4        19.5        18.9       15.0       14.2        10.1       20.0        14.3
172   •   Ratio and Valuation Worksheet



Risk Analysis
While modeling can help us calculate dollar value of an asset, the analyst’s kit bag
is incomplete without other tools such as industry knowledge and company
knowledge. Beyond operating performance and prospects, attributes of company
knowledge include management assessment, risk analysis, and financial strength.
A small subset of risk ratios analyzed across one or more cycles and in relation to
the market and peer group provides a solid base for determining financial
strength.
       Risk analysis ratios enable us to gauge a company’s financial strength in
relation to its past health, or lack thereof, and in relation to its peers. Some of
the ratios provide insight into a company’s ability to meet its cash obligations
while servicing debt. Others provide information on how a company’s operat-
ing profits will respond to a meaningful change in revenue levels. Some show
how encumbered a company is with debt in relation it its equity and its full
capitalization.
       Risk analysis ratios are a key determinant of financial strength. Calculating
financial strength via risk analysis ratios has its perils. The alternative is to rely
on credit rankings from the credit rating agencies—the same group that had
triple-A ratings on Fannie Mae and Freddie Mac nearly to the bitter end, and now
are in a (likely futile) race to fend off lawsuits with overly aggressive “impartial”
ratings on long-neglected corporate credits now deemed risky.
       For the analyst, a practical consequence of irresponsible M&A activity at
the company level and goodwill impairment at the FASB level is that not a few
companies have impaired away all (and more) of their stockholders’ equity. For
these companies, debt to cap is meaningless. In these situations, interest coverage
takes on disproportionate significance. And measures used in fixed-income anal-
ysis, such as measures of cash to indebtedness, take on new importance.


Operating Leverage
Operating leverage measures the magnitude of change in operating earnings in
relation to changes in revenue levels. As we discussed earlier, companies with
high fixed costs and capital intensity can be expected to have high operating
leverage, meaning that a rise (or decline) in revenue will produce a much larger
rise (or decline) in operating profits. This knowledge can influence investment
decisions, particularly at inflexion points in the economic cycle. While it is too
simple to say avoid high-fixed-cost companies heading into a downturn and buy
                     Ratio Analysis, Part 2: Return Ratios and Cash Flow Ratios   •   173



them heading into recovery, awareness of the effects of operating leverage on a
proposed investment should be part of the investment decision.
      To calculate operating leverage in a cell, you want to divide the percentage
change in operating income over the course of a period by the percentage change
in revenue over the course of the same period. Operating leverage calculations
are expressed as a percentage. For example, the one-period formula might look
like this: ((N102/M102) 1)/((N95/M95) 1), where M102 and N102 are operat-
ing income for successive years, and M95 and N95 are revenues for successive
periods.
      Single-period analysis is of limited use, in our view; to get more of a cycle-
spanning view, we calculate the five-period historical average for five years.


Times Interest Earned (Interest Coverage)
Times interest earned, sometimes called interest coverage or coverage ratio, mea-
sures a company’s ability to meet the accounting interest obligations on its debt.
While various operating earnings measures may be used, investors typically use
EBIT, or earnings before interest and taxes. Although EBIT, like net income, is a
book accounting concept, the exclusion of book taxes and book interest remove
the two inputs with the most variance from cash-based taxes and interest. There-
fore, it is the best measure to use.
      To calculate times interest earned in a cell, divide EBIT by interest. Recall
that we calculated EBIT along with EBITDA immediately below the income
statement.


Debt/Equity
The debt/equity ratio typically takes a back seat to the debt/capitalization
ratio. Debt/equity provides a quick snapshot on how a company chooses to
fund its strategy: with a preponderance of debt, with equity, or with a bal-
anced approach. Investors typically calculate both long-term debt/equity and
total debt/equity.
      To calculate debt/equity in a cell, divide period debt by period equity, both
brought up from the balance sheet. For this exercise, as for debt/capitalization,
you’ll need to sum all debt sources, including notes payable, short-term debt,
long-term debt due within one year (all to be found within current liabilities),
and long-term debt from long-term liabilities. Separately, to calculate debt/long-
term equity in cell, divide long-term debt only by stockholders’ equity.
174   •   Ratio and Valuation Worksheet



Debt/Capitalization
Like debt/equity, debt/capitalization (debt/cap) provides a snapshot of a com-
pany’s financing strategy. More so than debt/equity, investors use debt/capitaliza-
tion as a key ratio in the assessment of financial strength. Capitalization includes
all a company’s financing sources, including debt, stockholders’ equity, and
minority interest.
      A few things to keep in mind are that high debt/cap ratios can be prob-
lematic but also that debt/cap is really most useful in a peer group comparison.
Certain stable industries with high capital intensity, such as utilities, are char-
acterized by high debt/cap ratios. Other industries, including much of technol-
ogy, are characterized by generally low debt/cap ratios. Context is everything
with this ratio.
      We typically measure both total debt to cap along with long-term debt/cap.
These figures are often close. When there is a meaningful gap between them, that
usually signals maturing debt moved to current. If there is a meaningful gulf
between total debt/ cap and long-term debt/cap on a sustained basis, however,
that may indicate that a company is straining to operate profitably with its work-
ing capital resource. Such companies may be overly reliant on commercial paper
and/or may have nearly drawn down the totality of available revolving credit.
      To determine the debt/total capitalization ratio in a cell, sum all debt inputs
as above: notes payable, short-term debt, long-term debt due within one year (all
to be found within current liabilities), and long-term debt from long-term liabili-
ties. Divide this sum by stockholders’ equity plus total debt. To calculate the
long-term debt/capitalization ratio, divide long-term debt by the sum of stock-
holders’ equity and long-term debt.
      Few companies have been as resolute in regaining control of their financial
structure as Corning. In the midst of crisis in 2002, as bandwidth prices and fiber
demand collapsed, chairman Jamie Houghton returned to the CEO post and
made restoring blue chip finances his first priority. Figure 9.2 shows Corning’s
success in restoring financial stability, expressed in times interest earned and
debt/cap.


Cash Flow Ratios
In the next chapter we’ll move on to comparables valuation. But before we do, we
will take some time to compile a set of ratios directly related to cash flows. Our
model, remember, is not only constructed to deliver information; it is designed
Figure 9.2

Despite Corning’s success in improving its financial structure, the company’s operating leverage can be misleading. A substantial portion of income (from the
Samsung Corning Precision joint venture) does not appear in the operating income or pretax income lines and thus does not impact operating leverage.

Corning Inc.: Ratios, Valuations, Finances

Risk Analysis                    2000         2001        2002         2003       2004        2005      2006      2007        2008       2009E      2010E
Operating Leverage (1-yr)               0.1      (7.2)           1.0    (106.7)        17.3       4.2       2.2         2.7      (5.4)       10.0         2.7
Operating Leverage (5-yr)                                                            (19.1)    (18.3)    (16.4)     (16.0)         4.2        2.8         2.4
Interest Coverage                     5.5        (41.0)     (17.2)        (5.0)         0.8       5.6      12.5        17.3       19.2        6.3         9.9
Debt/Equity Ratio                  38.5%        91.1%      88.8%         51.5%      70.5%      32.2%     23.7%      16.2%       11.9%      13.2%      13.4%
LT Debt/Equity Ratio               37.3%        82.3%      84.5%         48.8%      58.0%      31.9%     23.4%      15.9%       11.4%      12.6%      12.9%
Debt/Capital Ratio                 27.8%        47.7%      47.0%         34.0%      41.4%      24.4%     19.1%      13.9%       10.7%      11.6%       11.8%
LT Debt/Capital Ratio              27.2%        45.2%      45.8%         32.8%      36.7%      24.2%     19.0%      13.8%       10.2%      11.2%       11.4%
176   •   Ratio and Valuation Worksheet



to facilitate organization, convenience, and information availability. These cash
ratios will mainly be used later in our discounted free cash flow analysis. But,
taking advantage of our stacked-up annual financial states, this is the best place
to situate them. And, as we collate cash flows even as we prepare to analyze sev-
eral earnings-based valuation techniques, this is as good a spot as any to weigh
in on the merits and demerits of each.


Cash Flows and Earnings
The broad mass of investors have come to rely on P/Es because they are ubiqui-
tous, easy to construct, and easy to comprehend. Many investors distrust earn-
ings-based valuations and instead have come to trust cash flows. A buck is a buck,
this thinking goes; cash talks and everything else walks.
       But cash flow is something of a straw man, put up (worse yet) by the earn-
ings gang. The very first input in cash flow from operations is net income, the
sup of the P/E crowd and the bane of the cash flow brigade.
       In advance of pursuing a CFA, in the mid-1990s, I took a course taught by
Paul Sondhi, one of the authors (along with White and Fried) of the excellent The
Analysis and Use of Financial Statements, (Wiley, 1994). He distinguished between
a “direct” cash flow statement and an “indirect” cash flow statement. The indi-
rect cash flow statement, which begins with net income, is an industry standard;
you see it in 10-Q and 10-K (respectively, the quarterly and annual filings man-
dated by the Securities and Exchange Commission for every publicly traded U.S.
company).
       One of our exercises in the course was to recast an indirect cash flows state-
ment into a direct cash flow statement. In the direct cash flow statement, every
line item is a summary of cash-based activities only: cash received from custom-
ers, rents, interest income, and so on, and cash paid for supplies, interest costs,
assets, and so on. Some casual investors may assume their cash flow–based mod-
els begin from a hard dollar basis. In fact, the cash flow–based models favored by
investment professionals all begin with the indirect cash flow statement.
       While cash flow modeling is thus subject to some of the same artificial
accounting constructs that (arguably) distort earnings-based models, it does
avoid some noncash inputs that can have unpredictable effects on earnings. As a
first step, cash flow excludes depreciation by adding it back to earnings. Depre-
ciation is logical; assets have useful lives. But the pace of decline in an asset’s
useful life rarely lines up exactly with the straight-line depreciation used in cal-
culating cost of goods sold.
                      Ratio Analysis, Part 2: Return Ratios and Cash Flow Ratios   •   177



      In real-world asset analysis, the key distinction between earnings-based
valuation and cash flow–based valuation lies less in their calculation differences
and the “veracity” of inputs, and more in their application time frame. For earn-
ings-based analysis (i.e., P/Es and relative P/Es), investors tend to use a relatively
lengthy (e.g., five-year) historical period to project value over a relatively short
forward period of one or two years. Cash flows, with fewer variables, better lend
themselves to adjusting or discounting to present value. Discounted cash flow
analysis is able to use a relatively short history to estimate (and discount back to
present value) cash flows over extended periods, up to and including a “terminal”
asset value calculation.
      Discounted cash flow valuation has been around since immediately after
the Great Depression. Pioneers in the field include Irving Fisher, who wrote The
Theory of Interest in 1930 (New York: the MacMillan Company), and John Burr
Williams, who wrote The Theory of Investment Value in 1938 (Harvard University
Press). In the modern era, discounted cash flow valuation has become wide-
spread because of the theory organization and refinement, advocacy, articulation,
and (unavoidable in our millennium) marketing efforts of Aswath Damodaran,
Distinguished Professor at the NYU Stern School of Business.


Operating Cash Flow
Immediately we need some clarity on terminology. As we’ll explain later, because
we rely on valuations of cash flow to equity (rather than cash flow to the firm),
we are in pursuit of operating cash flow for our analysis. The term operating cash
flow sounds an awful lot like cash flow from operations (CFO), the first and most
closely studied section in the cash flow statement. However, to project forward
operating cash flows, it is cumbersome to include every component of CFO,
many of which are one-time in nature.
      For this exercise, we use the term operating cash flow to distinguish it from
cash flow from operations, or the first section of the cash flow statement. We do
so, even though (in a further complication) some investors equate the term oper-
ating cash flow with EBITDA. For us, operating cash flow is the bare bones of
cash flow.
      While it is straightforward to use historical cash flow statements to assess
historical cash flow–based calculations, the value of cash flow analysis lies in
assessment of forward periods. It is arguably irresponsible to introduce additional
difficult-to-forecast elements into a calculation in which risk to the outcome is
extrapolated and exacerbated by the number of inputs and time’s passage.
178   •   Ratio and Valuation Worksheet



      Our definition and modeling of operating cash flow, by contrast, relies on
just a few inputs. Net income is the first; we’ve already modeled it with some
precision. The other inputs are depreciation and amortization. These inputs can
be reliably modeled, based on the fact that the underlying accounts (PP&E and
intangibles) are subject to scheduled, straight-line reductions. As well, the changes
in these accounts over time provide good guidance as to their levels in forward
periods. Finally, companies routinely provide guidance on their forecast depre-
ciation schedules for the coming year. No such guidance is furnished on any
other line item in CFO, with the possible exception of stock option compensation
(categorized in CFO as, for example, noncash compensation).
      To calculate operating cash flow in a cell, combine net income (from the
annual income statement) and depreciation and amortization (from the cash
flow statement) for the one-year period.


Free Cash Flow (FCF)
Within the cash flow statement, cash flow from operations is succeeded by cash flow
from investing (CFI) activities. While there is no hard and fast rule, in a normally
functioning enterprise the positive CFO number is succeeded by a negative number
in CFI. Routinely, the biggest single negative or cost in CFI is capital spending, rep-
resented by one of several names: expenditures PPE, capital costs, and so on. This line
item does not contain major asset purchases and is somewhat predictable in nature,
thanks to the operations and maintenance (O&M) component. Capital spending,
like depreciation, is a line item that companies often forecast in their addresses to the
investment community. Other line items in CFI are more erratic in nature, more
difficult to extrapolate, and not addressed in company guidance.
      Thus, in our cooked-down cash flow valuation model, we first adjust oper-
ating cash flow by excluding capital spending. This tally is universally called free
cash flow. To calculate free cash flow in a cell, adjust the operating cash flow
calculated in the cell immediately above by subtracting capital spending. Because
capital spending is presented in the cash flow statement as a negative number, in
practice your cell formula will reflect the addition of operating cash flow and
capital spending.


FCF Minus Change in Working Capital
Working capital, as discussed previously, represents the available near-term
resources a company may draw on without accessing the capital markets.
                      Ratio Analysis, Part 2: Return Ratios and Cash Flow Ratios   •   179



Although only the cash component is fully liquid, in the normal course of busi-
ness, accounts receivable will be paid and inventories will be converted into items
sold. As balance sheet items, current accounts and current liabilities and thus
working capital represent a reservoir of cash in varying states of availability.
      Remember, our concern is not with cash snapshots but with cash flows.
Thus we are less concerned about the actual amount of cash in the reservoir and
more concerned about changes in the cash reservoir across periods. Intuitively,
when working capital rises from one period to the next, cash is tied up and
unavailable to flow, and when working capital declines, the cash dispersed is part
of the cash flow.
      We thus need to adjust our free cash flow (FCF) further by the change in
the working capital reservoir. To calculate FCF minus change in working capital
in a cell, adjust the free cash flow calculated in the cell immediately above by
subtracting the change between prior-period-end working capital and current-
period-end working capital. To illustrate, the formula will look something like
this: =N39 (N4 M4), where N39 is current period free cash flow, N4 is current
period working capital, and M4 is prior-period-end working capital.


Free Cash Flow Minus Change in Trade Working Capital
So far we’ve included in our cash flow calculations the largest and most predict-
able sources and uses of cash. In our working capital discussion above, we identi-
fied three other line items that can represent the tying up or dispersal of significant
amounts of cash: accounts receivable, inventories, and accounts payable. We fur-
ther identified these three as the components of trade working capital.
      Always in cash flow valuation, we seek to limit the inputs to only those
accounts that significantly affect cash and, equally important, occur with suffi-
cient regularity to have a predictive component. The elements of trade working
capital, which meet this test, constitute their own reservoir of cash. To paraphrase
from what was said previously, when trade working capital rises from one period
to the next, cash is tied up and unavailable to flow; when trade working capital
declines, the cash dispersed is part of the cash flow.
      To calculate free cash flow minus change in trade working capital in a cell,
adjust the free cash flow calculated in the cell immediately above by subtracting
the change between prior-period-end trade working capital and current-period-
end trade working capital. You’ll recall that we calculated trade working capital
in the line immediately below working capital. To illustrate, the formula will look
something like this: N39 (N5 M5), where N39 is current period free cash
180   •   Ratio and Valuation Worksheet



flow, N5 is current period trade working capital, and M5 is prior-period-end
trade working capital.


Dividend Payout Ratio
Dividend payout ratio is a percentage representation that shows how much the
dividend payout is in relation to earnings. Net income is contributory to retained
earnings, but only after it has been modified by a few inputs; typically the chief
input is dividends paid.
      This percentage ratio is useful for a few reasons. In and of itself, it does not
show that much cash is being consumed in dividends and thus unavailable to
contribute to cash holdings. But as a percentage, it does show the trend in cash
consumption by dividends in relation to earnings. If dividend payout ratio is ris-
ing steadily in the normal course of business, then the dividend may well be at
risk in a down cycle.
      Separately, dividend payout ratio will contribute to retained earnings ratio,
which is a component of the constant growth rate used in discounted free cash
flow analysis. To calculate dividend payout ratio in a cell, divide dividends per
share by fully diluted GAAP earnings per share. Remember that we placed per-
share dividends at the bottom of our income statement presentation and then
carried them over to our annual income statement on the R&V page.
      Alternatively, you can use dividends paid from the annual cash flow state-
ment and GAAP net income from the annual income statement. To calculate
dividend payout ratio in a cell by this method, remember that dividends paid will
be represented as a negative number. And so you must adjust your formula to
something like this:      N190/N110.


Retained Earnings Ratio
The retained earnings ratio is also sometimes called the retention ratio. Histori-
cally, one measure of a firm’s progress in a normally functioning economy has
been growth in its stockholders’ equity account. Even with all our caveats about
the perils to stockholders’ equity—including management recklessness about
mergers and acquisitions and the fact that down cycles now routinely trigger
severe goodwill impairments—we still want to know how much and how regu-
larly a company plows its earnings back into the firm.
      The calculation for retained earnings ratio is net income minus dividends
divided by net income. Alternatively, and because retained earnings is the oppo-
site of dividend payout ratio, we can use the dividend payout ratio calculated
                     Ratio Analysis, Part 2: Return Ratios and Cash Flow Ratios   •   181



above to calculate retained earnings ratio. To calculate retained earnings ratio in
a cell, use the formula 1 N42, where N42 is dividend payout ratio.


Constant Growth Rate
Also called retained earnings growth rate, and often called by the shorthand g,
the constant growth rate provides information on the real rate of growth in
retained earnings. We mentioned in the preceding section that we wanted to
know the percentage of earnings being returned to the firm, shown by the
retained earnings ratio. Of even more intense interest is the underlying growth
rate of the firm.
      To calculate constant growth rate or retained earnings growth rate in a cell,
multiply return on equity by retained earnings ratio. Retained earnings ratio is
available immediately above, while ROE is a few rows higher in our ratios section
on our R&V page.
      Why adjust ROE by retained earnings? You could argue that return on
equity has implicitly been distorted and overstated by the exclusion of dividend
cost from retained earnings. The use of the retained earnings ratio to modify
constant growth rate is effectively an adjustment of that “distortion.”
      While changes in this number are useful for multiperiod analysis, this fig-
ure will be used extensively in our discounted free cash flow valuation. The
tumult in 2007 and 2008 reminded investors that any single year figure can be
distortive to the valuation process. Hence, for constant growth rate, we like to
calculate a five-year average rate to get a better picture of ongoing operations.
Figure 9.3 shows Corning’s strong cash generation patterns—and, because the
cash flow calculation begins with net income, it captures the strong equity income
component not reflected in Corning’s operating leverage (see Figure 9.2).


Finishing the Ratios Section
The ratios we have calculated and compiled above are far from exhaustive. We
think they represent a useful set of ratios that have informational value in and of
themselves; they can be assessed in comparison to industry peers and will con-
tribute to our valuation calculations. In time, each investor and analyst will
modify this list to his or her liking by adding some new ratios and discontinuing
or deemphasizing some on the original list.
     So far, all ratios have been calculated in a single column. They have been
constructed to lend themselves to drag-and-drop propagation across the page. To
Figure 9.3

Corning, a 150-year-old company, had to suspend its dividend amid the technology implosion in 2001–2002. The company reinstated the dividend in 2007, a posi-
tive move that nonetheless reduced its constant growth rate.

Corning Inc.: Ratios, Valuations, Finances

  Cash Flow                      2000         2001        2002        2003       2004        2005        2006       2007        2008       2009E       2010E
Interest Coverage                       5.5      (41.0)      (17.2)      (5.0)         0.8         5.6       12.5        17.3       19.2         6.3         9.9
Cash Flow (NI + DDA)                 1,122     (4,530)     (1,345)        428        1,183       1,115     2,456      2,770        3,473      2,379       2,940
Fre Csh Flw (NI+DDA-CX)                146     (6,271)     (1,702)          62        326       (438)       1,274     1,509        1,552      1,480       2,042
FCF - Chng in WC                   (2,110)     (5,644)     (1,788)      1,066         522     (1,137)         439     1,206        1,767      1,200        2,371
FCF - Chng in Trd WC              (2,253)      (4,745)     (1,399)         473        698       (804)         254     2,265        1,424       1,107       2,016
Dividend Payout Ratio                 19%         71%          0%          0%          0%          0%         0%         7%         13%         17%         14%
Retained Earnings Ratio              81%          29%        100%       100%        100%        100%       100%        93%          87%        83%         86%
Constant Growth Rate                   3%        -30%        -42%         -2%         -9%        11%         26%        21%         18%        11%          15%
5-Yr Constant Growth Rate                                                         -15.9%      -14.4%       -3.3%       9.3%       13.3%       17.3%       18.1%
                     Ratio Analysis, Part 2: Return Ratios and Cash Flow Ratios   •   183



expand these ratios across the page, drag and drop this column across the page
(preferably from left to right) for all historical periods and for all estimated
periods.
      All ratios that include two periods from the balance sheet have been designed
to encompass the preceding and current period, not the current and forward
period. If we used that current-and-forward-period construction, we’d drag and
drop to a meaningless conclusion for our final forward period—arguably the
most significant in setting stock value! So as a final test prior to the drag-and-
drop phase, test that your ratios drawing on multiple periods use past and
current.
      Keep in mind that, should you drag and drop to the left (meaning back in
time) and reach the final input historical column, any ratios that include two-
period averages will seek a prior period that is not included. You can exclude this
period from your analysis, assuming it is sufficiently distant from the five-year
comparison time frame. Alternatively, you can adjust this column to one-period
inputs; for example, assuming 1999 is the last historically input year, inventory
turnover for 1999 becomes that year’s COGS divided by that year’s inventories.
      Figure 9.4 shows the complete ratios section for Analog Devices Inc. We had
to shoehorn in an expanded risk analysis section when this company, long debt
free, decided to take on debt during fiscal 2009 (calendar 2008). As a dividend-
paying stock, the company now has a retained earnings ratio, but its five-year
constant growth rate signals a strong return on its capital structure. The com-
pany’s cash cycle seems kind of long; however, that partly reflects low inventory
turnover partly resulting from high gross margins as well as a strong overseas
customer base. (Overseas terms tend to be longer.) Overall, the ratios section
paints a picture of Analog Devices Inc. as an efficient, high-margined, and finan-
cially strong entity.
      That’s it for our extensive but not exhaustive ratios compilation. In the fol-
lowing chapter, we’ll use our modeled financial statement data and ratio analysis
to calculate value based on comparable historical valuation.
Figure 9.4

The complete ratios section for Analog Devices Inc. After a debt-free period in the middle of this decade, ADI now has some debt along with a manageable 13.2%
debt/capitalization ratio.

Analog Devices
RATIOS                                 1998          2002        2003       2004        2005        2006        2007         2008        2009E       2010E
  Internal Liquidity
Working Capital                             696         3,141      2,422       2,926       2,862       2,520        1,431       1,521       2,002       2,077
Trd Wrkg Captl                              479          443         483         549         597         568          517         455         391         405
Current Ratio                                4.4           7.5        6.2         6.2         5.1         6.1         3.6          3.7         6.6         6.7
Quick Ratio                                  2.5          6.5         1.8         0.6         0.5         0.7         0.6          0.6         0.7         0.7
Receivables Turnover                        12.1           7.6        7.8         8.4         7.0         7.5          7.4         7.9         7.1         8.8
Receivables Collection Period                30            48          47         43          53          49           49          46          52           41
Working Capital/Sales                        0.6          1.8         1.2         1.1         1.2         1.0         0.6          0.6         1.0         0.9
  Operating Efficiency
Inventory Turnover                             4.7        2.9        3.1          3.4        2.9          2.9         2.6         3.0         3.0         3.3
Inventory Turn Time (days)                     78        126         117         107        128          126         139         120         122          111
Days Sales Outstanding                        108        173        164          150        180           175        188         166          173        152
Fixed Asset Turnover                           3.5        2.0        2.8          3.9        3.9          4.6         4.4         4.6         3.7         4.4
Equity Turnover                                2.2        0.6        0.7          0.7        0.6          0.7         0.9         1.1         0.8         0.8
Payables Turnover                                       20.0        21.5        23.3        19.4         19.8        16.1        15.1        13.0        16.7
Payables Days Outstndng                                    18         17           16         19           18         23          24           28          22
Cash Cycle                                               155        147          135         162          157        165         142          145        130
    Return Ratios
Gross Margin                              47.8%       53.0%       54.9%       59.0%        57.9%      58.5%        61.2%       61.1%       55.0%       57.8%
Operating Margin                          12.8%        6.9%       18.3%       26.9%        21.6%      21.4%        23.1%       24.2%       12.1%       18.2%
Figure 9.4 (continued)


Analog Devices
RATIOS                      1998        2002       2003       2004       2005        2006       2007        2008       2009E       2010E
Pretax Margin                 12.2%        8.2%      18.7%      28.2%      24.6%       25.7%      26.8%       25.8%       12.7%       18.0%
Net Profit Margin               9.7%        6.2%      14.6%      22.0%       17.4%      21.3%      20.5%       35.4%       10.4%       13.9%
Return on Total Capital       21.2%        3.7%       9.6%      16.4%      11.1%       15.5%       17.5%      38.5%        8.4%       11.2%
Return on Assets              12.8%        2.1%       6.6%      13.2%        9.1%      13.1%      14.5%       30.2%        6.5%        9.1%
Return on Equity              21.2%        3.7%       9.6%      16.4%      11.1%       15.5%       17.5%      38.5%        8.4%       11.2%
EBITDA Margin                 22.6%       22.2%      26.9%      34.0%      31.1%       32.4%      33.1%       31.5%      20.4%       25.2%
EBITDA/Enterprise Value        6.0%        3.2%       4.5%       5.1%        5.3%       6.7%        7.0%       9.9%        5.0%        7.0%
Enterprise Value/EBITDA         16.5        31.2       22.4       19.5        18.9       15.0        14.2       10.1        20.0        14.3
  Risk Analysis
Operating Leverage                        (0.59)      13.21      20.51     (43.11)      56.46    (107.24)     105.67      (5.56)      25.52
5-Year Operating Leverage                                                    (2.8)        4.6       (9.7)        2.4         4.9         8.2
Debt/Equity Ratio             27.5%       43.9%       0.0%       0.0%        0.0%       0.0%        0.0%       0.0%        0.0%        0.0%
LT Debt/Equity Ratio          27.5%       43.9%       0.0%       0.0%        0.0%       0.0%        0.0%       0.0%        0.0%        0.0%
Debt/Capital Ratio            21.6%       30.5%       0.0%       0.0%        0.0%       0.0%        0.0%       0.0%       13.2%       11.3%
LT Debt/Capital Ratio         21.6%       30.5%       0.0%       0.0%        0.0%       0.0%        0.0%       0.0%       13.2%       11.3%
   Cash Flow
Cash Flow (NI + DDA)              247       286        464         721       568          716        639       1,063         358        462
Free Csh Flw (CF-CapX)             80       229        396         574       483         587         497         916         206        305
FCF - Chng in WC                (616)        (5)      1,115         71       546         928       1,587         826       (275)        229
FCF - Chng in Trd WC           (399)         171       357        509        434          616        548         979         269         291
Dividend Payout Ratio          0.0%        0.0%       0.0%      12.2%      27.0%       37.5%       47.1%      42.9%       95.8%       75.2%
Retained Earnings Ratio      100.0%      100.0%     100.0%      87.8%      73.0%       62.5%      52.9%       57.1%        4.2%       24.8%
Constant Growth Rate          21.2%        3.7%       9.6%      14.4%       8.1%        9.7%        9.2%      22.0%        0.4%        2.8%
5-Yr Cnstant Grwth Rate                                                    14.5%       13.1%      12.4%       11.3%        9.6%        9.5%
This page intentionally left blank
                                               Chapter 10
           HISTORICAL
           COMPARABLE VALUATION




Overview
The use of historical comparable ratios waxes and wanes, though P/Es are always
highly popular. The difficulty with historical comparables in general, including
P/Es, is translating the informational content into an investment decision based
on estimated dollar value of the asset. When we read, for instance, that ABC
Corp. is trading at 15 times earnings and usually trades closer to 20 times, that
sounds like a good deal. But do we have enough information to proceed? What
about cycle effects on past and current projected P/Es?
      Despite the shortcomings in P/Es, investors rely on them because they are
ubiquitous, easy to construct, and easy to comprehend. As mentioned, the market
moves not on realities but on shared perceptions. In the financial world, no per-
ception is as widely shared as P/E. So whether you like them or not, you’d better
understand—and calculate—P/Es.
      Fortunately, a range of other price-based historical valuations are available
for comparable analysis. In most industries, revenues are more stable than earn-
ings, even in down cycles. The inclusion of depreciated and amortized inputs in
cash flows tends to compensate for cyclical declines in earnings, resulting in
smoother cash flows for use in historical comparable analysis. Price/book value

                                                                               187
188   •   Ratio and Valuation Worksheet



is inherently stable, notwithstanding all the assaults on stockholders’ equity (and
all the ink we’ve spilled on that theme).
      There are even workarounds to normalize P/Es. Recall that we calculated
normalized earnings and described methods for calculating smoothed (ordinary
least squares) growth rates. Those inclined to do so can substitute these inputs
into forward P/Es.
      As a reminder, the key distinction between earnings-based valuation and
cash flow–based valuation lies in their application time frame. For earnings-
based analysis (i.e., P/Es and relative P/Es), a five-year historical look-back is used
to project value over the coming one or two years. Discounted free cash flows
value is predicated on determining present value of future cash flows; combined
with a terminal value for the asset, these are used to determine the dollar value
of the asset. The challenge with historical comparables is also to determine dollar
value of the asset, but perhaps not so far out in the future.


A Range or a Pinpoint Number?
Generating dollar value of an asset from comparables is a common and some-
what straightforward process. One such process is to determine the average for a
particular valuation technique and multiply that average by the forward input
per share used in the valuation methodology.
      For instance, to determine dollar value based on P/E, lets begin with a stock
trading at $20, a forecast of $1 per share in earnings, and a five-year average P/E
of 25. Multiply the five-year average P/E (25) for a stock times its earnings per
share ($1), and that provides a dollar value of $25. Given that $25 is 25% higher
than $20, the stock would appear to be undervalued.
      What I’ve just described is nonstandard. The standard methodology
on determining dollar value of an asset from historical comparables is more
conservative and, frankly, mushy. Analysts and investors are taught to cal-
culate the average five-period high valuation and average five-period low
valuation and then use them to forecast a price range in which the asset
should trade.
      I was taught this technique. It immediately became evident that the infor-
mation rendered by this technique provided almost no trading or investing guid-
ance. These low and high average inputs are based on the lows and highs in the
stock price. Intuitively, the low and high in the stock price stretch a few dollars
and maybe a few standard deviations from where they should be. When I used
average high and average low valuation, I almost invariably produced a valuation
range that captured 80% to 90% of the 52-week price range.
                                             Historical Comparable Valuation   •   189



      So, if you perform this analysis on a $21 stock and then tell the investor, “I
think the stock should trade in a range of $12 to $30,” why did you bother open-
ing your mouth? Even if you offer this same opinion and range on a $14 stock—in
other words, one much closer to the range low end and thus implicitly attrac-
tive—you’ve introduced such a wide range of potential prices as to mute the value
of your call.
      Yet at the same time we acknowledge the danger of using a single-point
valuation such as we illustrated with that $20 stock noted previously. To mini-
mize this danger, we increase the number of inputs. And to minimize the reliance
on an input as inherently volatile as earnings, we include more stable inputs, such
as revenues, book value, cash flow, and free cash flow.
      The point in common for all these inputs is that they can be calculated on
a per-share basis. We can use any per-share input; in some of our models, though
not all of them, we use EBITDA per share. For P/Es, price to sales, price/book, or
any input, the basic methodology is the same. Calculate the input per share per
period, divide that period input per share by the period price, calculate the five-
year average, and multiply that figure by the forecast period input.
      You can actually use any value converted to a per-share basis. For example,
you can calculate forward value using not just EBITDA but the percentage value
for EBITDA/EV. To do so, calculate an average stock price to EBITDA/EV per-
centage for each period, determine the five-year average, and multiply by the
forecast of period price divided by EBITDA/EV.
      We can use this valuation methodology, but we don’t. It is not proven,
understood, or accepted in the market. And in the world of historical compara-
bles, we only use what is proven, understood, and accepted.


Time Frames for Comparables Valuation
We all know that you calculate P/E by dividing the stock price by earnings. We
also know you calculate a trailing P/E by using most recent 12-month earnings.
And you calculate a forward P/E by using forward-period earnings.
      Does that mean that at midyear, your forward P/E estimate should borrow
the final two quarters of the current year and the first two quarters of next year?
It should. But this is one method that is honored in the breach. When investors
and analysts speak of forward P/Es, they are almost always using a full period’s
earnings rather than a bifurcated approach borrowing some from this year and
some from next year. More often, however, trailing 12-month P/Es really are
based on trailing four quarters.
190   •   Ratio and Valuation Worksheet



     Our goal is always to respect the convention or risk creating misleading
valuations. Unfamiliarity with the market’s not always consistent conventions
can result in misconceptions.
     Excluding the analysis we perform on December 31 and January 1, at any
other time we are amid periods. The convention is to have at least one full period
ahead on which to base valuation. In practice, the valuation target is both fluid
and soft. Thus, on March 7, 2002, investors were concerned about forward P/Es
for 2002 and 2003 but most focused on the 2002 P/E. By June 2002, they were
beginning to give more weight to the 2003 P/E. And by October 2002, they had
looked past 2002 P/Es, were most interested in 2003 P/Es, and were awaiting
some color on 2004 P/Es.


Lagged, Coincident, or Leading?
The question of valuation time frame raises another interesting question/
dilemma: should comparable valuations be lagged, coincident, or leading? Here
we are on the horns of the dilemma. Knowing that the market is an anticipatory
beast, investors tend to look forward—that is, they are much more concerned
with forward P/Es, despite the risk represented by modeled earnings, than with
historical or trailing P/Es, which can be calculated with unerring accuracy.
      We have touched on the calculation of the five-year historical periods used
as a basis to value forward based on current inputs. More than one budding ana-
lyst has asked: if we used forward P/Es in the valuation and purchase decision,
why use historical valuations that match the period rather than reflect the his-
torical anticipation that drove the asset purchase? In other words, why match
1996’s prices to 1996’s earnings, when 1996’s prices were based on anticipations
formed in mid-1995?
      It’s a fair question, and not easily answered. One way to solve the question
would be to adjust for, say, a half-year of anticipation. That is, in measuring his-
torical P/E for, say, 2003, use six average prices for calendar 2003 but use six
months of earnings from the back half of 2002 and six months from the front half
of 2003. We could even do it right on our ratios and valuation (R&V) page with
a few adjustments.
      The problem is that there is no optimal bifurcated period; we used only the
half year in our example because it’s, well, six of one, a half dozen of the other.
To better mimic investor thinking, why not seven months from 2002 and five
from 2003? Or four from 2002 and eight from 2003? In practice, the broad mass
of investors conducting comparable valuation analysis simply line up calendar
                                             Historical Comparable Valuation   •   191



year earnings (or fiscal year earnings) with calendar year prices (or fiscal year
prices). Once again your mother’s nightmare—everyone’s doing it, so you have
to as well—is required to attain true comparables compatibility outside and even
within your model.
      At the same time, we urge modelers to test and experiment with lagged and
leading inputs in performing comparables analysis.


Preparing the Historical Comparables Grid
For simplicity’s sake, we use the term “comparables” to stand in for comparables
historical valuation. We will situate this section beneath our ratios bloc and above
the annual stack of income statement, balance sheet, and cash flow statement. We
required about 40 to 50 lines on our R&V worksheet for our ratios analysis. We
will need a comparable amount of space for the comparables historical valuation
section.
      The P in P/E is price; per-share price is required for all the historical com-
parables. We already have current prices for the individual asset and the market
benchmark (the S&P 500), courtesy of our query worksheet. Our periods are
always, unless stated otherwise, one year. We need period prices for all historical
years. Ideally, we want high price, low price, and average price for each historical
year, with an emphasis on average price.
      There are numerous ways to gather historical prices. Pay services such as
Bloomberg and Thomson Reuters provide an average of all closing prices for the
approximately 250 trading sessions in any given year. If you lack access to a paid
provider, you can download prices for any security from Yahoo!Finance on the
handy Excel sheet provided right on its Historical Prices tab. Using the download
function right on the historical prices sheet, you can then use Excel to calculate
period average, and data sort to grab high and low prices.
      Or, for the generalist, you can sum the high and low price, divide by 2, and
call that the average for the year. Once again I’m baiting the purists. Yet we find
that here in the real world, the simple average of high and low falls within 90%
of the actual average nearly every time. I’m not advocating this shortcut, but I
never want the valuation process sidetracked by time constraints in a world of
jangling phones, harassing bosses, and piddling considerations like eating and
sleeping.
      We’re going use the price/earnings ratio as a means to describe the full
process of calculating comparable historical valuation and using it to project the
dollar value of the asset. Along the way we’ll discuss some challenges particular
192   •   Ratio and Valuation Worksheet



to this most integral of valuations. Most of the techniques described apply to all
other historical comparable valuations as well.
      Historical and current asset-price data will be used in all the comparables
valuations in our grid. For all of the comparables, we need two metrics. We need
the input (earnings, revenues, cash flow, etc.) on a per-share basis; and then we
need to divide that per-share input by the average price. We’ll then multiply the
five-year average of that value times the forward or estimated input to arrive at a
forecast asset value based on that information set.


Price/Earnings Ratio: Historical
As a start, beneath the valuations heading, prepare three more headings in
descending order—(ticker) stock price high, (ticker) stock price low, and (ticker)
stock price average—in three consecutive lines. Fill out the multiperiod range of
high, low, and average annual prices for the asset for at least a five-year period.
For both the current year and next year, input the current price from the query
page; we’ll explain why in a minute.
      In the line immediately beneath these three lines, input the appropriate
earnings per share directly below the appropriate period. Remember that as a
safeguard, we’ve built our income model to include both pro forma and GAAP
earnings. Input the appropriate measure that the market consensus uses to value
the stock. Some investors stand on principle and insist on the use of GAAP mea-
sures. You are doing your investors a disservice if you vary from the accepted
metrics used to value the stock in the market.
      We recommend entering three lines here: high P/E, low P/E, and average
P/E. Calculate high P/E by dividing (ticker) stock price high by period EPS. Cal-
culate low P/E by dividing (ticker) stock price low by period EPS. You can calcu-
late average P/E by averaging high P/E and low P/E or—our preferred method—by
dividing (ticker) stock price average by period EPS. Comparables models were
originally built with a high, low, and average price and a high, low, and average
for every valuation input (price/sales, price/book, etc.). This partly reflects the
habit of producing a valuation range rather than a pinpoint value. It may also
reflect the fact that in the past, obtaining full-year average price was cumbersome
and not easily available, and analysts often used the high-low average as a stand-
in for actual average. In any event, we rely on high and low valuation less and less,
and rely on true average valuation more and more.
      Repeat this process for every historical period.
                                             Historical Comparable Valuation   •   193



Price/Earnings Ratio: Forward
Historical P/E could not be more straightforward: period average price divided
by period earnings, producing a price/earnings valuation. Remember that our
R&V grid contains historical years as well as at least two full or partial forward
years: the current year and next year. What stock price do we use for forward
periods?
      We have several choices. If the current year is well along, we can use an
average price for the year to date. Or we can input the up-to-the-minute price,
which you’ll recall is on our query sheet. For the forward year, we can adjust the
current price to reflect the normal progress in the market. According to Ibbotson,
the acknowledged benchmark provider in these matters, the broad market has
historically delivered total average annual return in the 8% to 10% range. To
adjust the next-year price for “normal” market gains, you would multiply the
current price by 1.08 to 1.10.
      Personally, we tend to use current price, direct from the query page, for
both current year and next year. Typically, this is, if not the most accurate, then
the least inaccurate of all alternatives. Let’s shoot them down one at a time. If for
this year we use the average price year to date, we may be injecting past events
(earnings misses, executive firings, etc.) already discounted in the stock price
into a purportedly current valuation. If for next year we adjust for normal market
gains, then we introduce timing issues. The Ibbotson averages are most accurate,
intuitively, on January 1, a time when we’re wearing funny hats and swinging
noisemakers. Much more common is a look forward at next year’s price in March
or July; do we introduce another adjustment to our adjustment factor each time
we look at the stock?
      The main reason we use current prices in forward P/E calculation is that
when investors want a P/E, they assume your calculation has up-to-the-minute
price information. We again (and again and again) run into the need to conform
to the consensus methodology, given that the market runs on shared wisdom (or
shared hallucinations, depending on the degree of your dyspepsia).


Producing Dollar Value of the Asset from Comparables
We’ve now calculated P/E for all the historical years and for the current and
following year. The historical look-back is typically five years. In the line imme-
194   •   Ratio and Valuation Worksheet



diately below average P/E, in one cell identify the appropriate five-year period
(i.e., average 2004–2008); a few cells to the left, average the P/Es from 2004
through 2008.
       Create two more columns immediately to the right of the current year and
following year. Replicate the current and following year headers in these col-
umns. For simplicity’s sake, let’s say the current (2009) and following year (2010)
annual stack-ups are in columns N and O. We replicate the current year (2009)
and following year (2010) headers in columns P (2009) and Q (2010). These are
now the columns in which we’ll calculate dollar value of the asset. Furthermore,
let’s say our average 2004–2008 P/E has been calculated in line 53.
       In cell P53, multiply the 2004–2008 average P/E by forecast 2009 EPS. In cell
Q53, multiply the 2004–2008 average P/E by forecast 2010 EPS. These values tell
us the level at which the asset should be priced based solely on the relation
between historical average P/E and forecast earnings.
       Here are a couple of visual checks. Remember that we calculated P/E for the
current and following year. If our calculated P/E for 2009 is lower than the 2004–
2008 average, the forecast asset value should be higher than the current price. If
our calculated P/E for 2009 is higher than the 2004–2008 average, the forecast
asset value should be lower than the current price.
       Figure 10.1 shows Broadcom’s price history, pro forma earnings, and single-
year average P/Es for 2002 through 2008. It also shows the five-year (2004–2008)
average P/E of 24.7. Based on modeled earnings for 2009 and 2010, the P/Es on
those modeled earnings, and a current price of $28.54 imported from the query
page, in columns N and O Broadcom appears, respectively, overvalued on the
2009 pro forma earnings forecast and undervalued on the 2010 EPS projection.


Price/Sales Ratio
We now have two inputs to begin our calculation of dollar value of the asset based
on historical comparables valuation. Let’s gather a few more.
      Investors like the price/sales (P/S) ratio—and some prefer it over P/E—
because revenues are inherently more stable than earnings. We know that operat-
ing leverage—the degree of variation in operating income caused by variation in
revenue—impacts different companies differently, depending on their structure
and their operating execution. Revenue will vary, particularly within the eco-
nomic cycle, by typically much less. (An exception would be the end of bubble
periods, such as the Internet implosion in 2001–2002). Intuitively, greater stabil-
ity in a series increases the validity of the information contained in the historical
period calculation.
Figure 10.1

Broadcom built its EPS steadily from mid-decade through 2008, before EPS came crashing down in 2009. As such, the 2009 P/E climbed above the five-year aver-
age, indicating a value of $22.35 at a time when the stock traded above $28.

Broadcom

VALUATIONS                  1999       2002       2003         2004       2005      2006       2007      2008      2009E      2010E
BRCM Stock Price High                     24.88     24.88          31.07    33.00     50.00      43.07     29.74     28.54       28.54
BRCM Stock Price Low                       8.23      8.23           17.00   18.32      21.98     26.38     13.64     28.54       28.54
BRCM Stock Price Average                  16.56     16.56          24.03    25.66      37.59     32.62     21.73     28.54       28.54
EPS                            -         (0.31)      0.32            0.82    0.97       1.35      1.21      1.69       0.91       1.24   2009       2010
Average P/E                             (53.24)     51.95          29.47    26.37      27.86     26.90     12.83      31.52      22.96
                                                            Average 04-08             24.69                                                22.35      30.68
196   •   Ratio and Valuation Worksheet



      The historical and current asset price data we’ve gathered previously will be
used in all the comparables valuations in our grid. For all of the comparables, we
need two metrics. We need the input (revenues, cash flow, etc.) on a per-share
basis; then we need to divide that per-share input by the average price for each
period. Similar to our work with P/Es, we’ll then multiply the five-year average
of that value times the forward or estimated input to arrive at a forecast asset
value based on that information set.
      For the full price/shares calculation, we’ll need three lines: one for revenue
per share; one for average P/S; and one line for our five-year average P/S and
calculated asset value based on historical and forward inputs. For P/S, begin by
dividing period revenue by period diluted shares. These are conveniently avail-
able in our income statement stack for the period. We’ll work on one year and
then drag and drop for the historical and forward periods. In Figure 10.2, we see
that in year 2002, Alabama-based network equipment producer ADTRAN Inc.
(ADTN) had revenue of $346 million on a share base of 76 million diluted shares.
Revenue per share roughs out to 4.5 times. The average price per share of stock
in 2002 was $11.82. Therefore, we can say that for 2002, ADTN traded at an aver-
age price/sales ratio of 2.61.
      To calculate the dollar value of the asset based on P/S, we repeat the process
used with P/Es. In the third line of our P/S section, we calculate the five-year
average P/S. For consistency with our P/E discussion above, columns N and O
represent 2009 and 2010, respectively; column P represents the column in which
we calculate the dollar value of the asset based on 2009 inputs, while column Q
represents the column in which we calculate the dollar value of the asset based
on 2010 inputs.
      On the actual worksheet, row 60 shows period revenue per share, row 61
shows period average price/sales, and row 62 is where we compile our five-year
average P/S and calculate dollar value of the asset based on this input. To calculate
dollar value of the asset for 2009 based on P/S, in cell P62 multiply the five-year
average P/S times the 2009 average price to sales. To get that value based on 2010
inputs and current price, in cell Q62 multiply the five-year average P/S times the
2010 average price to sales. To ensure that your inputs are correct, you can use the
same visual checks you used for P/E. That is, if our calculated P/S for 2009 is lower
than the 2004–2008 average, the forecast asset value for 2009 should be higher than
the current price. If our calculated P/S for 2009 is higher than the 2004–2008 aver-
age, the forecast asset value should be lower than the current price.
      With two years of P/E and two years of P/S, we now have four inputs for
calculation of dollar value of the asset based on historical comparables valuation,
with more to come.
Figure 10.2

This illustration blends portions of the income statement with our comparable historical grid. The shaded cells all contribute to the average price to sales for 2001.
The five-year average of price to sales for 2004 to 2008, along with our forecast price/sales ratio for 2009 and 2010, suggest that ADTRAN is about 24% undervalued
at current prices.

ADTRAN
VALUATIONS                   1998           2002       2003        2004         2005      2006       2007       2008       2009E      2010E
ADTN Stock Price High           17.19          17.13     37.48        37.18       33.48     30.12      28.26      26.06      22.39       22.39
ADTN Stock Price Low             8.19           7.51     14.78       18.23        15.50     19.96      20.70      12.31      22.39       22.39
ADTN Stock Price Average        12.7          11.82      25.37       26.98        24.49     25.04      24.48       19.19     22.39       22.39

Revenue Per Share                3.66           4.52       4.91       5.62         6.58     6.28        6.85        7.77       7.50       8.45   2009E       2010E
Average Price/Sales              3.47           2.61       5.16       4.80         3.72     3.98        3.57        2.47       2.99       2.65
                                                                              Average 04-08             3.71                                        27.82       31.36

Income Statement                1998           2002       2003       2004         2005      2006       2007        2008      2009E      2010E
Net Sales                        287            346        397        455           513      474        477         501        472        535
COGS                             130             171       175        194           210      194        194         202        188        215
Gross Profit                      157            175        222        261          303       280        283         299        284        321

Basic Shares Outstndng             77            76          77         78          76         73         68         64          63         63
Diluted Shares Outstndng           78            76          81         81          78         75         70         64          63         63
198   •   Ratio and Valuation Worksheet



Price/ Book Value
Beware: Like many of the terms we use, book value has several meanings. In
general terms, book value is the value at which an asset is carried on the balance
sheet, reflecting any changes wrought on its value upon acquisition by deprecia-
tion, amortization, and impairments. For the enterprise, book value is the value
of stockholders’ equity. In equity valuation, book value is commonly regarded as
the per-share representation of stockholders’ equity. To clarify, in our valuation
discussion and analysis we use price/book value (P/BV) strictly to mean the rela-
tion of the stock price to the per-share value of stockholders’ equity.
      Beyond figuring in our comparable valuation analysis, price/book value
figures in the growth versus value debate waged by asset managers. Investors have
long held that stocks trading below the mean of some group (say, the S&P 500)
are value stocks, while those trading above are growth stocks. The two valuation
methods most often cited on either side of the mean are price/earnings and price/
book value; purists favor price/book value for the reasons we mention below.
      We have alerted investors to the perils to stockholders’ equity in the modern
era. Despite modern-era pressures on retained earnings, stockholders’ equity is
regarded as the best long-term record of a company’s earnings progress and as
the best record of the long-term management of the enterprise. Earnings for a
single period can oscillate because they have been deliberately or inadvertently
manipulated, because of unsustainable operating practices, because a fad or
mania supersedes long-term fundamentals, or some other factor.
      Calculation of dollar value of the asset based on P/BV is similar to the P/E
process and identical to the P/S process. For the inputs, we need the share count
for each period, available within our annual income statement stack, and stock-
holders’ equity at period’s end, available in our balance sheet stack.
      For the full P/BV calculation, we’ll need three lines: one for book value per
share; one for average P/BV; and one line for our five-year average P/BV and
calculated asset value based on historical and forward inputs. To calculate the
dollar value of the asset based on P/BV, we repeat the process used with P/E and
P/S. Drag and drop these single-period calculations to fill in values for all histori-
cal and forward periods. In the third line of our P/BV section, we calculate the
five-year average price/book value.
      Again using ADTRAN as well as example columns, lines, and cells,
ADTRAN closed 2002 with stockholders’ equity of $435.2 million. Based on the
76 million diluted shares outstanding that year, book value per share for 2002 was
5.69. The average price per share of ADTN stock in 2002 was $11.82; for 2002,
ADTN traded at an average price/book value ratio of 2.08. Dragging and drop-
                                              Historical Comparable Valuation   •   199



ping formulas across these two lines for all periods provides interesting informa-
tion, as we can see in Figure 10.3.
      Stockholders’ equity grew about 2.5% a year from 2001 to 2004; average
stock price grew faster, at a roughly 30% compound annual growth rate, and in
fact it more than doubled to $27 by 2004. Intuitively, this tells us that average
price/book value should be rising, and indeed for 2004 it was 4.8. For the 2004–
2008 period, price/book averaged 4.06, as the share price moderated to the mid-
$20s while stockholders’ equity continued its single-digit growth.
      At the point at which we performed this calculation, forecast price/book value
was 3.63 for 2009 and 3.20 for 2010. Based on our work so far, we’ve learned that
the lower P/BVs for 2009 and 2010 in relation to the five-year average P/BV will
result in a calculated dollar value for the asset higher than the current price, which
was about $22 at the time. And indeed, at the time of our calculation, our compa-
rables model produced P/BV values of $25.05 for 2009 and $28.47 for 2010.


Price/Cash Flow
Cash flow is an open-ended term with several meanings. In comparables valua-
tion analysis, we need a compact calculation of cash flow for projection purposes.
So in this case, we use the definition of cash flow to mean net income plus depre-
ciation and amortization. Conveniently, we’ve collated that above in our ratios
section. Instead of reaching down to get data from our annual income statement
or balance sheet stacks, we reach up to get the data from our ratios section.
      For the full price/cash flow (P/CF) calculation, we need one line for cash
flow per share, one for average P/CF, and one line for our five-year average P/CF
and calculated asset value based on historical and forward inputs. To calculate
dollar value of the asset based on P/CF, in our first line divide cash flow (calcu-
lated in ratios) by diluted shares. In our second line, divide average price by cash
flow per share. Drag and drop these single-period calculations to fill in values for
all historical and forward periods. In the third line of our P/CF section, we cal-
culate the five-year average price/cash flow.
      Figure 10.4 shows that for 2001, ADTRAN’s net income of $17.3 million and
depreciation and amortization of $16.8 million summed to $34.1 million. Based
on the 77 million diluted shares outstanding for that period, our simple calcula-
tion of cash flow per share was $0.44 (about twice the level of pro forma per-share
earnings). Based on the $12.23 stock price average for the year, P/CF for 2001 was
27.7. Cash flow and cash flow per share rose in subsequent years as earnings
strengthened while DDA and the share base stayed relatively constant. The five-
year average P/CF for 2004–2008 was 18.24.
Figure 10.3

In this illustration, we’ve left out the average pricing and per-share data furnished in Figure 4.9. But the same average price and share count inputs still figure in
our price/book calculation.

ADTRAN
VALUATIONS                    1998             2002       2003       2004          2005      2006      2007      2008       2009E      2010E


Book Value Per Share              2.95             5.69       6.12       5.77         6.95     5.79      5.44        5.84       6.16       7.01   2009E      2010E
Avg Price/Book Value              4.30             2.08       4.15       4.68         3.52     4.33      4.50        3.29       3.63       3.20
                                                                                 Average 04-08           4.06                                       25.05       28.47

Balance Sheet                 1998             2002       2003       2004          2005      2006      2007      2008       2009E      2010E
Cash & Equivalents              10.01            125.1      132.1      151.0         112.8      40.1      13.9      41.9       54.0       59.4
Shrt-Trm Investments            40.80             19.7       11.9       31.4         154.1      99.7     148.4      96.3        77.9      85.7

Common Stock                       0.4              0.4        0.8         0.8         0.8       0.8       0.9        0.9        0.9        1.0
Addtl PIC                         90.6             97.0      135.8       141.9       110.0     113.3     103.0      106.3      109.5      112.8
Accmltd OCI                          -              3.1       10.0        11.3         6.6       6.8       7.0        8.0        8.2        8.5
Retained Earnings                163.6            375.0      347.2      349.2        424.8     315.1     275.0      303.0      312.1      321.5
Treasury Stock                  (23.2)           (40.3)          -           -           -         -         -          -          -          -
   Shareholders Equity           231.4            435.2      493.8      466.9        542.2     436.1     378.4      375.8      388.0      443.7
   SE & Liabilties               301.7            521.2      593.9      559.8        652.6     542.4     479.2      473.6      495.0      553.6
                                             Historical Comparable Valuation   •   201



      Using our forecast inputs, on Figure 10.4 we calculated that P/CF was
17.89 for 2009—just under the five-year average—and 15.78 for 2010. Based
on what we have learned so far, intuitively this tells us that the dollar value
of the asset calculated with the 2009 inputs should be above the current price,
but not by much, and that the 2010 value should be comfortably above the
current price. And indeed, the value based on 2009 P/CF was $22.82, com-
pared with a price at the time of $22.39. The value based on 2010 P/CF, by
contrast, was just under $26.


Price/Free Cash Flow
As a reminder, simple free cash flow for comparable valuation purposes is simple
cash flow reduced by capital spending. We also calculated free cash flow in our
ratios section, and this line is also shown in Figure 10.4.
      Whereas cash-based depreciation for cash tax purposes is exponential,
“book” or accounting depreciation is scheduled as a simple average of depreciable
asset value, levied on a constant basis year after year. Assuming no massive
change in depreciable assets, depreciation does not move around much. Capital
spending, by contrast, can be highly variable. At the down cycle, companies
reduce capital spending to not much more than that required for operations and
maintenance spending as a means to conserve cash. At the time of our writing,
the company was preparing to follow a year (2008) of relatively high capital
spending with more subdued spending plans for the current and following year.
      In general, companies disclose their capital spending plans, particularly at
watershed moments in the economic cycle. They do so partly out of self-preser-
vation; no management wants to be seen as profligate with cash when core net
income is weak. Accordingly, the price/free cash flow valuation (P/FCF) is inher-
ently less stable than price/cash flow valuation.
      You know the drill by now. To calculate P/FCF, allocate one line for free
cash flow per share; one for average P/FCF; and one line for our five-year average
P/FCF and calculated asset value based on historical and forward inputs. To cal-
culate dollar value of the asset based on P/FCF, in our first line divide free cash
flow (available in our ratios section) by diluted shares. In our second line, divide
average price by free cash flow per share. Drag and drop these single-period cal-
culations to fill in values for all historical and forward periods. In the third line
of our P/FCF section, we calculate the five-year average price/cash flow.
      Intuitively, use of this input has several predictable outcomes. Free cash flow
per share will be lower than cash flow per share. P/FCF will be higher than P/CF.
The five-year average P/FCF will be higher than the five-year average P/CF.
Figure 10.4

This illustration blends portions of our ratio analysis, specifically devoted to deriving cash flow, with our comparable historical calculation of price to cash flow.

ADTRAN
RATIOS                       1997       2001        2002      2003       2004        2005      2006       2007      2008       2009E      2010E
  Cash Flow
Cash Flow (NI + DDA)                           34       41         77         91        117         95         93         96         79        90
Free Csh Flw (CF-CapX)                         21       39         70         84        108         89         87         87         72        83

VALUATIONS                   1997       2001        2002      2003       2004        2005     2006        2007      2008       2009E      2010E
Cash Flow Per Share                        0.44        0.54      0.95       1.13        1.50      1.26       1.34      1.49        1.25       1.42   2009E      2010E
Avg Price/Cash Flow                       27.72       21.94     26.68      23.98       16.28     19.84      18.24     12.84       17.89      15.78
                                                                                   Average 04-08                                            18.24      22.82      25.88
                                            Historical Comparable Valuation   •   203



      ADTRAN’s experience bears this out. In 2001, the company reduced its
$34.1 million in cash flow by $13.2 million in capital expenditures, resulting in
simple free cash flow of $20.9 million. Free cash flow per share for 2001 was 0.27,
fractionally above EPS, and P/FCF for the period was 45.24.
      Five-year average P/FCF for 2004–2008 was 19.74—about 8.2% higher than
the five-year average P/CF for 2004–2008. Based on our forecast inputs, at the
time of this exercise P/FCF was 19.50 for 2009—barely under the five-year aver-
age—and 17.05 for 2010. Based on what we’ve learned so far, intuitively this tells
us that the dollar value of the asset calculated with the 2009 inputs should be
above the current price by even less than the price premium based on 2009 fore-
cast P/CF. The value based on 2009 P/FCF was $22.66, or less than the $22.82
value based on 2009 P/CF. Value based on 2010 P/FCF was nearly $26.


Relative P/E
Relative P/E refers to the relation between the price/earnings of the company and
the market price/earnings. The market proxy is almost invariably the S&P 500,
a market-weighted index. The earnings of the 500 members of the S&P 500 are
also market weighted to determine the earnings of the index. The market P/E is
the price level of the S&P 500 divided by the annualized four-quarter earnings
of the index.
      Can we calculate dollar value of the asset based on relative P/E? Of course.
The process is similar to the process we’ve used with our other comparable valu-
ations but more complicated; it also needs a few more inputs and a few more lines
on our valuation grid.
      We’re going to allocate about 9 or 10 lines to the relative P/E section, and
we’re going to consider the calculation in two separate pieces. First we need a
place to calculate the historical and forward market P/E, along with its five-year
average. Next we need to calculate the relative P/E of the asset. Using this infor-
mation, we’ll calculate dollar value of the asset based on this relationship. Given
the (slightly more) complex nature of this calculation, refer to Figure 10.5 as we
go along.
      Start with three lines for S&P 500 high, low, and average price. You can
gather this information for the historical periods from the usual pay or free
sources. Historical S&P 500 prices are straightforward. For the forward periods,
what price is best? I recommend using the current S&P 500 price level for both
the current and following year. Recall the key points we made in our discussion
of the appropriate price to use for the individual asset in P/E and other compa-
rable valuation. Average S&P 500 price for the year to date may fail to discount
Figure 10.5

The interaction of the asset input (2009 forecast EPS), the market input (the 2009 market P/E based on forecast S&P 500 adjusted earnings), and the P/E relation-
ship between ADTN and the market (relative P/E for the five-year historical period) enable us to project dollar value of the asset based on relative P/E. Note that
the projected rise in S&P 500 earnings in 2010 from depressed 2009 levels makes ADTN less attractive based on 2010 inputs, even though the company’s earnings
are rising—just not as fast as S&P 500 earnings.

ADTRAN
VALUATIONS                    1998         2002        2003       2004         2005      2006       2007        2008       2009E      2010E
ADTN Stock Price High            17.19        17.13      37.48       37.18       33.48     30.12      28.26       26.06      22.39       22.39
ADTN Stock Price Low              8.19         7.51      14.78      18.23        15.50     19.96      20.70       12.31      22.39       22.39
ADTN Stock Price Average        12.70        11.82       25.37      26.98        24.49     25.04      24.48        19.19     22.39       22.39
EPS                              0.51         0.32        0.76       0.93         1.29       1.14       1.17        1.32       1.15        1.27   2009E     2010E
Average P/E                     24.68        36.47       33.30      28.97        18.94     21.93      20.86       14.54       19.50       17.61
                                                                             Average 04-08            21.05                                         24.16     26.76

S&P 500 High                  1,241.00      1,172.51   1,111.92     1,214       1,270.0   1,427.0    1,562.0     1,425.4      923.3      923.3
S&P 500 Low                     970.00       776.76     800.73      1,063       1,136.0   1,223.0    1,374.0       752.4      923.3      923.3
S&P 500 Average               1,085.50       993.76     965.23       1,131      1203.0     1310.0    1,468.0     1,088.9      923.3      923.3
Operating Earnings               45.37         48.61       55.4      66.5         77.00     87.75      82.50       50.00      50.00      63.00
S&P 500 Avg P/E                  23.93         20.44      17.44     17.00         15.62     14.93       17.79      21.78      18.47      14.66
                                                                             Average 04-08             17.43                             -21%

Average Relative P/E              1.03         1.78       1.91        1.70         1.21    1.47         1.17       0.67        1.06       1.20
                                                                             Average 04-08             1.25                                         26.39     23.20
                                               Historical Comparable Valuation   •   205



the new events reflected in the current price. Meanwhile, attempts to adjust next
year’s S&P 500 price level must accommodate a constantly moving time frame.
       In the next line record the historical annual earnings for the S&P 500; your
best source for this information may be the Standard & Poor’s Web site, although
it is available elsewhere as well. We run into a familiar question: for the S&P, do
we use reported earnings or adjusted earnings? We use adjusted earnings; the
valuation calculations for the 500 component stocks are most often performed
based on adjusted earnings. For the forward period, both top-down and bottom-
up forecasts of market earnings are widely available. (Top-down forecasts come
from market strategists and investment strategists; bottom-up forecasts represent
the consensus views of individual analysts.)
       Using this information, in our fifth line we can calculate average market
P/E for the historical periods as well as the forward periods. In a sixth line, we
always calculate the five-year average of market P/E. This number does not figure
directly in our valuation calculation. But every self-respecting analyst should
have an awareness of where the market traded for the past five years, and it’s nice
to have it at hand.
       The next step is to calculate the historical and forward relative P/Es for the
individual asset. To do so, divide the P/E for the asset by the market P/E. We need
just a single line for this exercise. Below this line, we include a line in which we
calculate the five-year historical relative P/E for the asset.
       In 2002, the market multiple was a 20.44, reflecting still overheated stock
levels and fast-retreating earnings. With its rich 36.47 average P/E multiple for
2002, ADTRAN traded at an average relative P/E of 1.78 for the year. Over the
2004–2008 period, as the market P/E came in, ADTRAN’s P/Es declined even
more meaningfully.
       To calculate the dollar value of the asset based on relative P/E, we need
input from all players: that is, from the asset, from the market, and from the
relationship between the asset and the market. For ADTRAN, we’ve been calculat-
ing the dollar value of the asset for 2009 and 2010 in columns P and Q, meaning
the two columns adjacent to N and O, where we are modeling forward valuations
for 2009 and 2010. For 2009, we’ll calculate the dollar value of the asset based on
relative P/E in column P.
       Of the three inputs specifically required in this exercise, the first is forecast
2009 EPS; that is available in our income statement stack but—most handily—in
our P/E calculation as well. For the market input, we’ll use the S&P 500 P/E based
on forecast 2009 earnings. And for the relationship between the two, we’ll use the
five-year average relative P/E. To calculate the dollar value of the asset, multiply
these three inputs together. For 2009, this formula tells us that on a relative P/E
206   •   Ratio and Valuation Worksheet



basis, the ADTN shares are worth $26.39, or about a $4 premium to the stock
price at the time of calculation.
       What are our checks on the validity of his method? Our 2010 calculation pro-
vides some insights. The calculated dollar value should increase or decline based on
changes in the inputs. The five-year average P/E for 2004–2008 is not going to change;
so the inputs that can vary are company EPS for the period and forecast market P/E
for the period. Moreover, it will be the net effect of changes in the two variables; a big
change in one should supersede a more modest change in the other.
       For 2010, our model suggests that ADTRAN will earn $1.27, up 11% from
our $1.15 forecast for 2009. But bigger changes are afoot in market valuation.
Both market strategists and individual analysts are anticipating something like a
normal—that is to say, robust—earnings recovery out of the recession. So the
consensus call is for S&P 500 earnings to rise from a net $50 for 2009 to $63 for
2010—a 21% growth rate. For 2009, ADTRAN looked “cheap” relative to the
market at a relative P/E of 1.06. If all our expectations prove out, ADTRAN will
be more expensive relative to the market in 2010, at a relative P/E of 1.20—very
close to its historical 2004–2008 relative P/E of 1.20. We would thus expect cal-
culated dollar value for ADTRAN based on relative P/E for 2010 to be below that
in 2009. At $23.20 for 2010, ADTRAN’s dollar value based on 2010 relative P/E is
12% below the calculated value of $26.30 for 2009. And that 12% rate of change
is in the ballpark of the difference between the 11% gain in ADTRAN’s forecast
earnings between 2009 and 2010 and the market’s 21% gain in forecast earnings
between 2009 and 2010.


PEG and PEGY
Recall that between P/E and P/S we left a considerable bloc of six to seven empty
lines. In this space, we want to calculate a set of values that do not contribute to
our comparables dollar-value calculations but that make sense being situated
near our P/E calculations. PEG and PEGY are two more values we could wrangle
into a per-share calculation and use to calculate dollar value of the asset; but we
choose not to. We think they provide more visual information value; they’ll also
find a place in our industry matrix.
      PEG is the ratio of P/E valuation to forecast or actual growth in earnings.
PEGY is the ratio of P/E valuation to total return, meaning earnings growth plus
dividend yield. These calculations are easy enough to prepare given the inputs
and calculations abounding in the valuation section, so this is where they are
situated. We can see ADTRAN’s PEG and PEGY in Figure 10.6; we can also see
that ADTRAN began paying a dividend in 2003.
Figure 10.6

The calculations of PEG and PEGY are not manipulated to estimate dollar value of the asset—although they could be. This information is most useful when compared
across the peer group.

ADTRAN
VALUATIONS                     1997      2001       2002       2003      2004         2005    2006       2007       2008     2009E     2010E
ADTN Stock Price High                      15.00       17.13     37.48      37.18       33.48    30.12     28.26      26.06    22.39     22.39
ADTN Stock Price Low                        9.00        7.51     14.78     18.23        15.50   19.96      20.70      12.31    22.39     22.39
ADTN Stock Price Average                   12.23      11.82      25.37     26.98        24.49   25.04      24.48       19.19   22.39     22.39
EEPS                                        0.22       0.32       0.76      0.93         1.29     1.14       1.17       1.32    1.15       1.27   2009E     2010E
Average P/E                                54.59      36.47      33.30     28.97        18.94    21.93     20.86      14.54    19.50      17.61
                                                                                    Average 04-08          21.05                                    24.16    26.76

LT EPS Growth                              12.0%     12.0%      12.0%     12.0%     12.0%    12.0%        12.0%      12.0%    12.0%      12.0%
PEG Ratio                                    4.55      3.04       2.77      2.41      1.58     1.83         1.74       1.21     1.63       1.47
                                                                                 Average 04-08              1.75
Dividend Yield                              0.0%      0.0%       0.6%      1.2%      1.4%     1.4%         1.5%       1.9%     1.6%       1.6%
PEGY                                         4.55      3.04       2.64      2.20      1.41     1.63         1.55       1.05     1.43       1.29
208   •   Ratio and Valuation Worksheet



       For the numerator in PEG and PEGY, we’ve already calculated P/E, and it is
immediately available to us. One-period earnings change, needed for short-term
PEG, requires nothing more than a line in which we measure the percentage change
in earnings growth. Forecast long-term earnings growth is available from pay or free
sources. Don’t forget that in our normalized earnings and regression (ordinary least
squares) section, we measured normalized and smoothed earnings growth; these
calculations may provide the best basis for forecasting long-term growth. To get the
yield for PEGY, remember that we carried the dividend payout over from the income
statement presentation onto our R&V page for just this sort of situation.
       So let’s carve out seven lines for our PEG and PEGY section. We don’t typi-
cally bother with short-term PEG, which is seen as having a lower quality of
investible information. But in case we want that information in the future, we
usually add a line in which we calculate the annual change in EPS on a percentage
basis. In our next line, we enter our forecast five-year growth rate. You can enter
the consensus forecast, or you can use historical normalized earnings or ordinary
least squares to produce your own five-year growth forecast.
       In the next line we calculate PEG ratio. For this calculation, we need to
divide a percentage (growth) into a number (P/E); so we have to “square up” the
calculation by multiplying growth by 100. On our ADTRAN R&V worksheet,
where P/E for 2001 is in cell F52 and long-term growth is in cell F57, the formula
will look like F52/(F57*100). Using a constant 12% long-term growth rate,
ADTRAN’s PEG ratio dropped from 4.55 early in the decade to 1.21 by 2008.
After dragging and dropping to derive PEGs for all periods, in the following line
calculate the five-year historical average. Again using a constant 12% rate,
ADTRAN’s five-year average PEG was 1.75. As a general rule, the lower the PEG
ratio, the more attractive the asset.
       For PEGY, we need to calculate dividend yield in a convenient spot. In the
next dedicated line, divide the period annual dividend (from the annual income
statement) by the period average price (from the top of the valuation grid).
ADTRAN began paying a slight dividend in 2003; by 2008 the yield was above
1.5%. Given that this dividend yield is additive to total return, PEGY for ADTRAN
should be lower (and more attractive) than PEG for the company. And indeed,
five-year average PEGY is 1.57.


Finishing the Grid: Dollar Value of the Asset Based
on Comparables
We have calculated two years’ worth of six inputs—P/E, P/S, P/BV, P/CF, P/FCF,
and relative P/E—that we will use to calculate dollar value of the asset. Note that
                                            Historical Comparable Valuation   •   209



these need not be the only inputs used in our comparables valuation. Any input
that can be rendered on a per-share basis can be included; many analysts will use
EBIT and EBITDA per share to calculate dollar value of the asset. As further
noted, means and methods exist in which we can tweak percentage ratios—such
as EV/EBITDA, PEG, or PEGY—into a per-share configuration and derive dollar
value of the asset on that basis. But the method is unproven; more importantly,
the market has no place for it.
      To complete the process, we will average the six values created in column P,
for 2009, and the six in column Q, for 2010. Finally, we will create an average of
these two values. At the time of our calculation, with the ADTN common shares
of ADTRAN trading just above $22, Figure 10.7 shows that our comparables
historical valuation process calculated a fair value for ADTN of $24.47. Just to
show it could be done, we have also included a series in which we calculate dollar
value of the asset based on enterprise value (EV) to EBITDA. Though shown in
the grid, this series is not included in our actual valuation work.


Annual Updates
Around the time you’re finishing your first ratio and historical comparable valu-
ation grid and feeling pretty good about yourself, you suddenly realize it’s Octo-
ber. The market is no longer looking at the nearly done year and next year; it is
looking at next year and the year after that. What to do?
       Based on the assumption that your income statement presentation has been
modeled to include that new year, let’s get right to it. Insert a column between
your last modeled year (2010 in this case); label that column 2011. Drag and drop
the 2010 column into the 2011 column.
       Within the annual income statement stack, you’ll now have meaningless
data in the 2011 column. Link the line items in this column to the actual full-year
2011 column on the income statement presentation. Prepare and adjust the 2011
balance sheet and cash flow statement estimates, either by growing the accounts
at a reliable multiple (which is the easy way) or performing a more complicated
interaction among all the financial statements.
       For your oldest modeled year (the current year, of which two months are
left in October), you can begin to think about substituting high, low, and average
price for current price; by year end, that data needs to go in. Ditto for the S&P
500 Index price.
       Move the five-year average time frame from 2004–2008 to 2005–2009. We
find the F2 function key highly useful at this time, because if you strike it from
within a cell it will show the components of any formula in this cell. From within
the five-year average cell for P/E hit F2; move the highlighted range down one
Figure 10.7

At the time this “snapshot” of the comparables historical valuation section was taken, the 12 inputs for historical comparable valuation for ADTRAN suggested the
stock was worth just under $24.50 per share, or about 9% more than its price at the time.

ADTRAN
VALUATIONS                  1997      2001       2002       2003      2004         2005     2006       2007       2008       2009E      2010E
ADTN Stock Price High                   15.00       17.13     37.48      37.18       33.48     30.12     28.26      26.06      22.39      22.39
ADTN Stock Price Low                     9.00        7.51     14.78     18.23        15.50    19.96      20.70      12.31      22.39      22.39
ADTN Stock Price Average                12.23      11.82      25.37     26.98        24.49    25.04      24.48       19.19     22.39      22.39
EPS                                      0.22       0.32       0.76      0.93         1.29      1.14       1.17       1.32      1.15        1.27   2009E     2010E
Average P/E                             54.59      36.47      33.30     28.97        18.94     21.93     20.86      14.54      19.50       17.61
                                                                                 Average 04-08           21.05                                       24.16    26.76

LT EPS Growth                           12.0%     12.0%      12.0%     12.0%        12.0%    12.0%      12.0%      12.0%       12.0%      12.0%
PEG Ratio                                 4.55      3.04       2.77      2.41         1.58      1.83      1.74       1.21        1.63       1.47
                                                                                 Average 04-08            1.75
Dividend Yield                           0.0%      0.0%       0.6%      1.2%         1.4%      1.4%      1.5%        1.9%       1.6%       1.6%
PEGY                                      4.55      3.04       2.64      2.20         1.41      1.63      1.55        1.05       1.43       1.29
                                                                                 Average 04-08            1.57
Revenue Per Share                         5.00      4.52       4.91      5.62         6.58      6.28      6.85       7.77       7.50       8.45
Average Price/Sales                       2.44      2.61       5.16      4.80         3.72      3.98      3.57       2.47       2.99       2.65
                                                                                 Average 04-08            3.71                                       27.82     31.36

Book Value Per Share                      5.66      5.69       6.12      5.77         6.95      5.79      5.44       5.84       6.16       7.01
Avg Price/Book Value                      2.16      2.08       4.15      4.68         3.52      4.33      4.50       3.29       3.63       3.20
                                                                                   Average                4.06                                       25.05     28.47
                                                                                    04-08
      Figure 10.7 (continued)


      ADTRAN
      VALUATIONS                1997   2001        2002       2003       2004         2005     2006       2007       2008      2009E      2010E
      Book Value Per Share                5.66        5.69       6.12       5.77         6.95     5.79       5.44       5.84      6.16        7.01
      Avg Price/Book Value                2.16        2.08       4.15       4.68         3.52     4.33       4.50       3.29      3.63       3.20
                                                                                    Average 04-08           4.06                                     25.05       28.47

      S&P 500 High                     1,373.00    1,172.51   1,111.92     1,214      1,270.0   1,427.0   1,562.0    1,425.4     923.3      923.3
      S&P 500 Low                        965.00     776.76     800.73      1,063       1,136.0  1,223.0   1,374.0      752.4     923.3      923.3
      S&P 500 Average                   1,194.18    993.76     965.23       1,131      1203.0    1310.0   1,468.0    1,088.9     923.3      923.3
      Operating Earnings                   45.79      48.61       55.4      66.5         77.00    87.75     82.50      50.00     50.00      63.00
      S&P 500 Avg P/E                      26.08      20.44      17.44     17.00         15.62    14.93      17.79     21.78     18.47      14.66
                                                                                    Average 04-08           17.43                           -21%

      Average Relative P/E                 2.09       1.78       1.91       1.70         1.21     1.47        1.17      0.67      1.06        1.20
                                                                                    Average 04-08            1.25                                    26.39       23.20

      Cash Flow Per Share                  0.44       0.54       0.95       1.13         1.50     1.26       1.34       1.49       1.25       1.42
      Avg Price/Cash Flow                 27.72      21.94      26.68      23.98        16.28    19.84      18.24      12.84      17.89      15.78
                                                                                    Average 04-08          18.24                                     22.82       25.88

                                                                                                                                                             (continued)
211
Figure 10.7 (continued)


ADTRAN
VALUATIONS                 1997   2001      2002      2003       2004        2005     2006       2007      2008      2009E      2010E
EV/EBITDA                           3.9%      6.5%      4.9%       5.7%        8.7%      6.8%       7.4%    11.3%       8.1%       9.1%
Avg Price/EV/EBITDA                317.54    181.91    520.37     477.32      282.01   367.75     328.77    170.31     276.19    246.44
                                                                           Average 04-08         325.23                                    26.37   29.55

Free Cash Flow Per Share             0.27      0.50      0.87       1.04        1.39      1.18      1.25      1.35       1.15       1.31
Avg Price/Free CF                   45.24     23.45     29.27      26.03       17.62     21.20     19.61     14.24      19.50      17.05
                                                                           Average 04-08          19.74                                    22.66   25.92

                                                                Comparables Value                                                          25.04   23.90
                                                                                                                                                   24.47
                                              Historical Comparable Valuation   •   213



year. Change the years in the adjacent title cell. Copy these cells and paste for
every series that requires a five-year average.
      The value calculations, once performed in columns P and Q, will have moved
over to columns Q and R because of the inserted column. Once again, use the F2
key to highlight formula components; adjust to the following year. Be particularly
careful when adjusting the relative P/E calculation, as it has extra inputs.


Historical Comparables for ADRs of Foreign Companies
We have been consistent in modeling ADRs in their native currency, and you
might assume this complicates the modeling process. Actually, the main head
scratcher with international companies that trade dollar-based ADRs is to decide
which discounting mechanism to use. There is no real template.
      When we originally began covering Nokia in the early 2000s, the dollar/
euro relationship was fairly close. The euro, which began at dollar parity and then
slipped to as low as $0.90, by 2002 was worth more than the dollar. For most of
that year, the euro was worth from $1.05 to $1.15.
      If we assume that this exchange rate is going to be stable at around $1.10,
and assuming a 1/1 ADR to share exchange rate, we can very simply build in a
9% to 10% P/E discount to compensate for foreign country risk. Assuming we
use our existing historical comparables grid and apply it to Nokia, we would show
Nokia’s ADRs priced in dollars. If we priced EPS in euros rather than dollars, the
P/E would reflect a roughly 10% discount to the P/Es priced in dollars; this 10%
discount would serve as a proxy for foreign country risk.
      This method began to break down in 2003, when the euro/dollar cross
moved from $1.06 to $1.23, and is completely out the window now that the euro
is near $1.50. We now adjust the currency as well as the inputs for every historical
comparable calculation immediately above the valuation calculation; we adjust
for the foreign country risk factor at a later stage in the valuation process.
      Take a look at our final illustration in this chapter, Figure 10.8. It shows our
translation work for Ericsson. (We considered using Nokia or Alcatel-Lucent, but
the euro/dollar cross is dynamic at present, and the latter company is finding
profits highly elusive.) We will focus on the shaded area. The first row, showing
the stock/ADR conversion, is an anachronism; it dates from the period in which
the ADR/common share ratio was 10/1. Now it is 1/1. The next line is most
important; here we convert krona to dollars. Note that for each year, the krona/
dollar conversion uses the average data directly from the income statement.
Figure 10.8

We can use our historical comparables valuation grid to value ADRs of overseas companies such as Ericsson even if those companies are modeled in their home
currency. The shaded section enables conversion first of the currency and then of the various financial statement inputs required.

Ericsson

Stock/ADR Conversion                           2,514.6     3,164.6    3,165.8      3,164.8   3,174.2     3,185.9   3,201.8    3,193.0    3,202.0
Kroner/Dollar Conversion                          0.104      0.124      0.136        0.133     0.136       0.153     0.152      0.127      0.134
EPS/ADR                                          (0.78)     (0.43)       0.82         1.02       1.12       1.06      0.59       0.65       0.82
Revenue/ADR                                     15,095     14,653     17,999        20,197   24,242      28,773    31,692     27,983     32,524
 SE/ADR                                           7,622      7,527    10,972       13,925    16,294      20,549     21,361    18,035      19,585
 Cash Flow/ADR                                 (1,292)       (305)      3,347        4,007    4,581       4,005      2,438      2,351      3,331
 Free Cash Flow/ADR                               (417)      (263)        122           85       102         339     (257)     (255)       (167)
 ERIC ShrPrice KR                               137.66       44.15       87.14     112.64    121.56      103.28      64.18      73.13      69.36

VALUATIONS                 1997      2001      2002       2003       2004          2005     2006        2007       2008      2009E      2010E
ERIC ADR Price High                              26.86        8.54     15.66         17.20    19.22       20.81      13.83      9.31        9.31
ERIC ADR Price Low                                 1.65       2.45      8.11         12.77    13.76       10.84       5.64      9.31        9.31
ERIC ADR Price Average                            14.26       5.50     11.89         14.99    16.49       15.83       9.74      9.31        9.31
EPS ($)                                          (0.78)     (0.43)      0.82          1.02     1.12        1.06       0.59      0.65        0.82   2009E    2010E
Average P/E                                     (18.21)    (12.89)     14.50         14.66    14.70       14.86      16.43     14.33       11.40
                                                                                 Average 04-08           15.03                                       9.76    12.28
      Figure 10.8 (continued)


      Ericsson

      Current Yr EPS Growth             -46%    -292%       25%      10%        -5%     -44%        10%      26%
      LT EPS Growth                    15.0%    15.0%     15.0%    15.0%      15.0%    15.0%     115.0%    215.0%
      PEG Ratio                        (0.86)     0.97      0.98     0.98       0.99      1.10      0.12      0.05
                                                       Average 04-08           1.00
      Revenue Per Share         6.00    4.63      5.69      6.38     7.64       9.03     9.90      8.76      10.16
      Average Price/Sales       2.37    1.19      2.09      2.35      2.16      1.75     0.98      1.06       0.92
                                                       Average 04-08            1.87                                 16.36       18.96

      Book Value Per Share      3.03    2.38      3.47        4.40     5.13    6.45      6.67      5.65       6.12
      Avg Price/Book Value      4.70    2.31      3.43        3.41     3.21    2.45      1.46      1.65       1.52
                                                         Average 04-08         2.79                                  15.77        17.08

                                                                                                                             (continued)
215
Figure 10.8 (continued)


Ericsson

S&P 500 High               1,373.00    1,172.51   1,111.92   1,214       1270.0   1427.0      1,562.0       1,425.4   996.31   996.31
S&P 500 Low                  965.00     776.76     800.73    1,063       1136.0   1223.0       1,374.0        752.4   996.31   996.31
S&P 500 Average             1,194.18    993.76     965.23     1,131      1203.0   1310.0     1,468.00     1,088.90    996.31   996.31
Operating Earnings             45.79      48.61       55.4    66.5        77.00     87.75        82.50        50.00    50.00    63.00
S&P 500 Avg P/E                26.08      20.44      17.44   17.00        15.62    14.93          17.79       21.78    19.93    15.81
                                                                      Average 04-08              17.43

Average Relative P/E                     -0.89      -0.74     0.85         0.94     0.98         0.84         0.75      0.72     0.75
                                                                      Average 04-08              0.87                                   11.30   10.78

Cash Flow Per Share                     (0.51)     (0.10)     1.06         1.27      1.44        1.26         0.76      0.74     1.04
Avg Price/Cash Flow                     -27.75     -57.05    11.24        11.84     11.43       12.59        12.79     12.65     8.95
                                                                      Average 04-08             11.98                                   8.82    12.46

Free Cash Flow Per Share
Avg Price/Free CF
                                                                                            Comparables Value                           12.40   14.31
                                                                                                                                                13.36
                                            Historical Comparable Valuation   •   217



      In the next few rows, we use the dollar/krona relationship to convert krona-
based inputs—EPS, revenue, stockholders’ equity, cash flow, and free cash flow—
that we will need in our standard comparable historical inputs of P/E, P/S, P/BV,
P/CF, and P/FCF. Note that we need not provide anything specific for relative
P/E, as the inputs noted have already captured what we need.
      As an example, consider price/sales for 2008. In the shaded area, we see that
based on the dollar/krona average exchange rate for 2008, Ericsson’s revenue
amounted to $31.7 billion in U.S. dollars. Using the stock/ADR conversion data
(again, ADR to common is currently 1/1, but we keep the converter available in
case the company again changes that ratio), we see that Ericsson had the equiva-
lent of $9.90 per-share in revenue in 2008, and that, based on an average ADR
price for that year, its average price/sales ratio was 0.98. Just as with our U.S.-
based companies, we compile the averages and model the forward inputs—all
translated into dollars—to arrive at the dollar value of the asset.
      In our view, the U.S. analyst had better be equipped to model foreign stocks,
at least those that trade ADRs on U.S. exchanges. The BRIC nations (Brazil, Rus-
sia, India, and China) are an investing obsession. Resource-rich BRAC nations
(Brazil, Russia, Australia, and Canada) are of growing interest to investors as
well. ADRs provide the opportunity for easy investment overseas; the currency-
adjusted model provides a common ground in which to situate the company.
      We’ve now drawn on financial sheet data and some of our ratios in the
calculation of historical comparable valuation. Other ratios and data points that
we’ve not yet used go to work in our next section, in which we discuss the most
common form of present value analysis known as discounted free cash flow valu-
ation. Like comparable historical valuation, present value and DFCF contain pit-
falls and dangers for the uninformed analyst—particularly given the degraded
state of stockholders’ equity. But also like historical comparables, present value
methodologies provide tremendous amounts of information useful to the
analyst.
This page intentionally left blank
                                                            PART 3

           STOCK VALUE
           WORKSHEET




In Part 3 of this book, we will dive into our next worksheet, stock value, which
is so called because all the valuation conclusions will coalesce here. Stock value
is the site of our present value calculation based on discounted free cash flow
(DFCF). It is also the place where we aggregate our other valuation inputs, such
as the dollar value estimate based on comparable historicals. We’ll also include
values based on our proprietary technique known as Peer Derived Value, although
that input needs to traverse a roundabout path: out to the industry matrix work-
book (our next task) and then back to the individual company workbook.
      Our first task on the stock value worksheet is to build the second of our
valuation methodologies, based on discounted free cash flow analysis. Our sec-
ond task is to use the stock value worksheet to compile and integrate three valu-
ation methodologies: (1) DFCF analysis and (2) comparable historical valuation,
two widely accepted methodologies, along with our proprietary relational valu-
ation methodology known as (3) Peer Derived Value (PDV). We’ll then risk-ad-
just the integrated valuation to estimate the dollar value of the asset, and—based
on its relation to current price—make our asset decision or our investment rec-
ommendation (most typically buy, hold, or sell).
      A word on content organization on this particular worksheet and in gen-
eral: most valuation texts focus on concepts. We’re allotting the concepts little

                                                                              219
220   •   Stock Value Worksheet



more than a thumbnail sketch while focusing mainly on the alignment of data
on the worksheet and across the workbook. That’s a necessary step when one is
concerned with organizational flow from one worksheet to another.
      We’ve been fairly specific about layout so far. The income statement presen-
tation is straightforward; it represents an expanded version of what the compa-
nies themselves provide. We modeled the income statement so the EPS would be
readily available for P/E calculation; we also needed to reference COGS for inven-
tory turns, dividends for retained earnings percentage, and so on. In our ratios
and valuations worksheet, we’ve planted annual information in straight furrows
in the field, the better to harvest the data; no data farmer wants a crazy-quilt
layout.
      Our stock value layout, which includes both our DFCF process as well as the
compilation of other valuation conclusions (comparables and PDV), is arguably
arbitrary. There’s no compelling reason for how we organize data on this work-
sheet. We recommend a specific layout mainly because it has always worked for
us. For that reason, we’re going to ask you to focus a bit more on the accompany-
ing illustrations in this section. You might want to examine them in detail and
then return to the text. As always, we suggest you do it our way at first; you can
rearrange the structure to your liking later on.
      In the beginning of Chapter 12, we show the stock value worksheet from
Cisco to illustrate the main components of our layout. In the upper left corner is
a subset of data points—some imported from other worksheets, some to calcu-
lated—that will be used for weighted average cost of capital calculations. On the
top right, we keep a running tally of return on equity calculated by the DuPont
technique.
      In the middle of the worksheet, beginning on the left, we perform dis-
counted free cash flow valuation analysis. Typically we do so in a two-stage model
and a three-stage model. We can perform the calculation of cash flows one of two
ways, and we’ll detail that in the following paragraphs.
      At the bottom left of the worksheet, we tally up various valuation inputs
from this page and from historical comparables. We also include an input from
a proprietary industry-specific valuation technique that we call Peer Derived
Value (PDV), which is imported back from our industry matrix workbook. We
average these inputs to determine a “blended” dollar value of the asset.
      Finally, at the bottom middle we have a section wherein we assess the dif-
ference between the current stock price (imported from the query page) and the
blended value. This difference shows the capital appreciation potential of the
asset. We’ll also add current yield (imported from the ratios and valuation page)
to get a sense of total return potential on a risk-adjusted basis.
                                                       Stock Value Worksheet   •   221



      Whether you’re an analyst assigning buy, hold, and sell ratings or a portfolio
manager making buy, hold, or sell decisions, you need to assess total return
potential in relation to the market benchmark. The benchmark, or S&P 500, has
a long history of delivering 10% average annual total return (unadjusted for
inflation). Over time, you’ll set your own risk bands. But for starters, you could
say that an asset whose total return potential exceeds the market average (i.e.,
10%) warrants a buy, an asset within the market total return band (0% to 10%,
or alternatively –10% to 10%) should be held, and an asset whose total return
potential is less than –10% should be sold.
      Before we make that call, however, we’ll risk-adjust the total return poten-
tial by adjusting for the asset’s variability from the market. We could use standard
deviation. In the realm of U.S. equities measured against the U.S. market bench-
mark, we use beta, and we’ll explain how in the following paragraphs.
      If you have some familiarity with DFCF schemes, you may be thinking that
risk adjusting with beta amounts to double counting. Beta is already factored into
the cost of equity (the capital asset pricing model, or CAPM), which is a key
component of the discount rate (for discounting to the firm) and the only dis-
count rate (for discounting to equity). We would argue that with the inclusion of
other value inputs (e.g., comparables and PDV) into blended value, beta is muted
nearly completely. Besides, the consensus risks-adjusts its return assumptions,
both implicitly and explicitly. Not only do asset managers make a like (i.e.,
explicit) adjustment, the investment style whereby riskier investors pursue riskier
assets is a form of implicit risk adjustment.
      Now that we have discussed the stock value worksheet methodology, let’s
proceed. In Chapter 11 we discuss the place of present value analysis in relation to
other accepted methodologies; we also touch on the advantages, disadvantages,
dangers, and benefits of discounted free cash flow, and why it is our preferred
present value methodology. In Chapter 12, we delve into DFCF theory and meth-
odology for the individual company. And in Chapter 13 we present two techniques
for deriving value of the asset from discounted free cash flow valuation.
This page intentionally left blank
                                                 Chapter 11
           PRESENT VALUE
           MODELING AND
           THE STOCK VALUE
           WORKSHEET




Present Value versus Comparables
Which is better, historical comparables or present value analysis? Even though
our fully realized architecture later on will indicate which technique we weight
more heavily, always remember that “which is better?” is rarely what the market
is asking. The market is ruled not by reality, but by perception, and specifically
by shared perceptions. The collective trust in P/Es and to a lesser extent in present
value calculations means that asset prices will respond to the shared perceptions
of value based on those valuation methodologies.
       Even if we thought that discounted free cash flow (by far the most widely
accepted present value methodology) and historical comparables had no more
basis in reality than phrenology, we would fold them into our valuation method-
ology for (as your mother always told you) the worst possible reason: everyone
else is doing it. Again, our goal is to drive cattle, not argue theory until the camp-
fire goes out. If the market is using it, we’re using it—and we’d be irresponsible
if we did not. Fortunately, we think these methodologies have value. And, with
precise modeling, we think we can add insight of our own.
       We employ both historical comparables and DFCF because we believe they
provide valuable information on asset valuation and the investment decision. But

                                                                                  223
224   •   Stock Value Worksheet



we are as mindful of their limitations as we are of their relative advantages; here’s
our short-hand on each.
      Use of historical comparables should invoke the disclaimer language on
your mutual fund proxies that warns that past performance is no guarantee of
future outcomes. Certainly, the five-year look-back on which we base compara-
bles provides insight into the present and future. But as the revolution in infor-
mation technology accelerates the pace of change at companies and whole
industries, that value may be limited and is certainly incomplete. Every cycle
leaves different detritus in its wake; some come through like a rainstorm, others
like Hurricane Katrina. The downturn of 2007–2009 has had structural and secu-
lar implications, in our view, that supersede the normal cycle effects. Depending
on industry and the damage sustained, five-year look-backs dating from 2004–
2008 may be as applicable in coming years, to paraphrase the writer Luc Sante,
as crinoline and moustache wax.
      Discounted free cash flow valuation methodologies are more forward look-
ing. The danger in looking forward is to utterly misread the future. The danger
in discounted free cash flow is subtly different; if we begin with incorrect or
misjudged inputs, with each new forward increment we amplify the original erro-
neous assumptions. Discounted free cash flow also directly incorporates more
market, economic, and interest rate assumptions than historical comparables,
which are mainly inward looking. It is tough enough forecasting growth rates for
a single enterprise; adding assumptions related to the market, the economy, and
interest rates adds to predictive risk.
      Nonetheless, and much as historical comparables have demonstrated their
validity, discounted free cash flow does provide demonstrated value. After all,
companies are dedicated to safeguarding and growing their cash flows, and it is
right to model the ongoing enterprise, because the enterprise is ongoing. If we
incorporate outside forces (e.g., risk-free rates) in our DFCF valuation, we must
be as rigorous in adapting our market expectations as we are in updating indi-
vidual company cash flow expectations. With these guidelines in mind, we can
jump into our abbreviated DFCF valuation methodology.


Thorns and Brambles of Discounted Free Cash Flow
Historical comparables is safely shrouded in the mists of time; its time-tested
concepts are no more contentious than a phone book. Discounted free cash flow
valuation is not exactly new, but it is newly propagated. Its adherents are as vocif-
erous as they are splintered; they’re ornery about their turf.
                         Present Value Modeling and the Stock Value Worksheet   •   225



      As is our custom, as much as possible we’ll circumnavigate theory and drive
the herd right past the academic literature. While it is unavoidable that we touch
on the key concepts underlying DFCF, we intend to touch lightly. No matter how
light the touch, of course, we express a preference.
      Be forewarned: we are about to discuss a very abbreviated form of DFCF
valuation. We structured our income statement presentation as a framework
enabling basic modeling with the proviso that it permits more detailed asset
modeling and analysis as the modeler becomes more familiar with the entity.
Here, too, we create a framework of basic modeling that can over time be adjusted
to encompass the specific circumstance of the asset. To delve into every nuance
of discounted free cash flow would risk sending the modeler down countless
byways in a quest to anticipate every special situation. We are all for taking on an
oceanic concept such as DFCF; we just don’t want to drown in the puddle of
details.
      We regard discounted free cash flow as another tool in the box rather than
the tool box itself. Some DFCF adherents find it contradictory to use both a pres-
ent value scheme and a historical comparables scheme. The market, however, is
not so pure; asset prices reflect the smorgasbord of inputs, from complex algo-
rithms to hot tips. While we haven’t decoded the market’s DNA valuation, we can
provide the modeler with a range of tools so as to have the right one at hand at
the right time.
      While reiterating our need to keep the cattle drive moving, we’ll take the
risk of driving the herd right into a DFCF border war by questioning a key
concept associated with DFCF modeling. The constant growth rate that we
modeled in Chapter 9—defined as retained earnings percentage times stock-
holders’ equity—is a frequently modeled growth rate in DFCF valuation.
Accordingly, we need both inputs to be rock solid—or at least reasonably con-
stant. We would argue that stockholders’ equity is under pressure on a variety
of fronts; furthermore, the industry does not appear to be very concerned with
reversing this trend.


Assessing Behavioral and Structural Impacts on
Stockholders’ Equity
Are companies today less “responsible” than they’ve been historically? Certainly,
the operating environment in which they play has changed. The field has leveled
to encompass players from all nations. Most new business originations are in low
capital intensity fields such as technology and business services; in the future,
outsourcing and “virtual infrastructure” enablers such as cloud computing could
226   •   Stock Value Worksheet



further lower capital needs to launch a company. Gary Wendt, former CEO of GE
Capital, once said that anyone with a phone, a yellow pad, and a pencil could be
a competitor to his firm; update this to the digital age, and a smart phone alone
may be enough.
       In capital-intensive areas, unprecedented capital flows and the demise of
capital borders has made $100 billion (and more) mergers and acquisitions
(M&A) deals a reality. The proliferation of hedge funds and a concomitant explo-
sion in high-velocity trading has telescoped the performance analysis period for
many investors from an already too-tight 90 days to something much less.
       With competitive barriers lowered to admit one and all, competitive threats
to the entrenched are rising. Companies may feel a need to move dangerously far
beyond core competencies in pursuit of excess return. In this environment, risk
taking in pursuit of improved returns has become not just common but expected.
CEO behavior once deemed prudent is now seen as stodgy and a ticket to early
retirement. When risk taking is not just rewarded but requisite, the outcome can
be big and unproven M&A bets as well as the displacement of tactical adjust-
ments by strategic lurches.
       With all these corporate reinventions going on, it follows that the number of
failed corporate experiments must rise as well. Risen they have, at least in part
because the penalty for failed experiments is itself devalued. Given investors’ will-
ingness to trade on adjusted rather than GAAP results, failed strategies come crash-
ing through GAAP results—and are forgotten. The “Big Bath” was at one time
orchestrated to coincide with the market down cycle; even so, this bundling of bad
news cloaked the executive suite in a mantle of shame. Now, serial restructurers
blithely report operating fiascos related to strategic missteps every quarter.
       Even the appropriately strategic and perfectly executed acquisition, how-
ever, will likely wrack the balance sheet and thus stockholders’ equity over time.
In our opinion, the Financial Accounting Standards Board, with its implementa-
tion of FAS 142, doomed companies to a nearly inevitable barrage of nonrecur-
ring charges in the form of goodwill impairment. As usual we can blame good
intentions, which pave the road to Wall Street and other destinations.
       Goodwill used to be depreciated. In the go-go 1990s, pressure was brought
to bear on FASB to stop depreciating goodwill and instead subject it to periodic
tests. The thought, in those “New Paradigm” days, was that depreciating goodwill
was a tax on earnings, and anyway the market was going up forever. According
to FASB Summary of Statement No. 142, issued in June 2001, the substituting of
impairment of goodwill for depreciation of goodwill “will improve financial
reporting,” because financial statements from acquiring companies will “better
reflect the underlying economics of those assets.”
                         Present Value Modeling and the Stock Value Worksheet   •   227



       FASB mandates impairment tests on a regular basis. When the underlying
value of like assets declines, as in a broad market collapse, goodwill associated
with those like assets must be impaired to mirror the market experience.
       Almost as soon as this change was enacted early in the decade, the market
collapsed. Huge piles of goodwill that previously would have been worked down
gradually had to be impaired all at once. For a decade or more, technology compa-
nies had been printing stock certificates, watching investors bid them to hysterical
heights, and using them to buy other helium-infused assets. When the crash came,
technology companies were forced to practice impairment by chain saw. JDSU and
Nortel were two companies that purchased vastly inflated assets by issuing vastly
inflated stock. In 2001, Nortel reported a $24 billion loss, most of which came from
goodwill impairments. But that is peanuts compared with JDSU, which in its fiscal
2001 year reported a $45 billion loss, again mainly from impairments. Nortel as we
went to press was nearing extinction. JDSU has been able to get back to a positive
shareholders’ equity account, mainly because its additional paid-in capital account
(also called capital surplus) of $69.4 billion at year-end fiscal 2009 was roughly
equivalent to its retained deficit of $69.5 billion.
       Nontech investors in the early part of the decade snickered at technology’s
profligate ways. There’s no one left to laugh now; the market evisceration of
2007–2009 (and counting) has spared no sector. Amid a market in collapse,
acquisitions once deemed as prudent in 2005 or 2006 have been suddenly revealed
to carry the baggage of overvalued goodwill—all of it to be whacked at once,
regardless of future market prospects.
       The switch to impairing rather than depreciating goodwill can lead to spec-
tacular hits to earnings in down cycles, which in turn flow through and reduce
retained earnings and thus stockholders’ equity. This heartier corporate appetite
for strategic risk, and the concomitant impact on earnings based on failed strate-
gies, takes a further bite out of stockholders’ equity.
       To test the validity of this perception, in Figure 11.1 we measure S&P 500
GAAP earnings per share as a percentage of S&P 500 “adjusted” earnings per
share, going back to 1989 in two 10-year periods. Whereas operating earnings
represented 90% of GAAP earnings between 1989 and 1998, the operating earn-
ings to GAAP percentage is below 78% for the period 1999 to 2008.
       As expected the ratio has degraded over time, with the decline intensifying
in the more recent 10-year period (much of which was subsequent to the change
in goodwill accounting). While two years in the more recent decade contributed
disproportionately to this discrepancy, that is precisely the point. The years in
which the gulf between GAAP and operating earnings is most glaring are 2003
(59.9%) and 2008 (30.1%). These are years of major market declines in which
228   •   Stock Value Worksheet


Figure 11.1

Riskier acquisition activity, the FAS decision to impair rather than depreciate goodwill, and investors’
willingness to ignore GAAP losses and focus on ongoing operations have all weakened the ties
between GAAP earnings and adjusted earnings.

S&P 500 Earnings: GAAP vs. Operating

                            OPERATING                AS REPORTED
                            EARNINGS                  EARNINGS
                             PER SHR                   PER SHR                     GAAP EPS
                             (ests are                 (ests are                 as a Percentage
          Year              bottom up)                top down)                   of Oprtng EPS
          2008                         49.51                         14.88                      30.1%
          2007                         82.54                         66.18                     80.2%
          2006                         87.72                         81.51                     92.9%
          2005                         76.45                        69.93                       91.5%
          2004                         67.68                        58.55                      86.5%
          2003                         54.69                         48.74                      89.1%
          2002                         46.04                         27.59                     59.9%
          2001                         38.85                        24.69                      63.6%
          2000                         56.13                        50.00                       89.1%
          1999                         51.68                         48.17                     93.2%
                                                          10-Year Average                      77.6%

          1998                          44.27                        37.71                       85.2%
          1997                          44.01                        39.72                       90.3%
          1996                          40.63                        38.73                       95.3%
          1995                          37.70                       33.96                        90.1%
          1994                          31.75                       30.60                        96.4%
          1993                          26.90                        21.89                        81.4%
          1992                          20.87                        19.09                        91.5%
          1991                          19.30                        15.97                       82.7%
          1990                          22.65                        21.34                       94.2%
          1989                          24.32                        22.87                       94.0%
                                                          10-Year Average                        90.1%

Source: Standard & Poor’s




companies were compelled to impair goodwill regardless of future prospects for
the acquired businesses.
      To further test this hypothesis, we looked at the ratio of stockholders’ equity
to market capitalization over an extended time frame. The implication of our
thesis is that stockholders’ equity should be shrinking as a percentage of market
capitalization if stockholders’ equity is being decimated by flawed strategic think-
                         Present Value Modeling and the Stock Value Worksheet   •   229



ing, and, equally important, investors are not treating this as a grave event. Our
research supports the view that stockholders’ equity has declined as a percentage
of market capitalization.
      Despite our overheated rhetoric, we’re not condemning this shift so much
as reporting it because of its long-tailed impact on the valuation process. Return
on equity (ROE) is one of the most closely monitored measures of corporate
performance. Impairments and restructuring charges reduce stockholders’
equity; but, as noncash events, they do not degrade the ongoing enterprise. A
company with stable earnings that has reduced its stockholders’ equity through
these events will suddenly report much-improved return on equity.
      The seeming counterbalance to this development would be that decrements
to stockholders’ equity would cause expansion in debt/capitalization ratios. Obvi-
ously, if equity is reduced and debt is unchanged, debt/cap will rise. Many bonds
carry indentures that permit the debt to be called should debt/cap ratios exceed
certain limits. These should act as gatekeepers to undue risk in corporate strategy
and M&A activities. Yet in the past decade, many of these indentures have been
rewritten to exclude effects from impairments to stockholders’ equity based on
noncash events. In other words, another mechanism that might have served to
limit risky strategic behavior has been defanged.
      While both ROE and debt/cap may be sending false valuation signals in the
immediate term, our chief concern is not with ROE on a stand-alone basis. ROE is
a linchpin metric in the calculation of valuations based on discounted free cash
flow. The constant growth rate, sometimes called g, is calculated as retained earn-
ings percentage (calculated as earnings per share less dividends per share, divided
by EPS) multiplied by ROE. If ROEs, as we are suggesting, are at risk of being over-
stated, than forward growth prospects are at risk of being overstated. In that situ-
ation, the risk would be in the assessment of the long-term value of the asset.
      Again, we’re not here to bang the podium about wrong or right. Our sole
concern is that this behavioral shift in risky corporate behavior may be causing
distortions in valuation measures based on return on equity. We’ll tackle this
topic in more depth in Chapters 12 and 13 during our discussion of discounted
free cash flow valuation. For now, we’ll just say that investors need to be aware of
the sensitivity of valuation analysis to the inputs to stockholders’ equity.


Dividend Discount Model and the Financial Structure Decision
Why have we embraced one present value methodology and disowned another?
     Discounted free cash flow is the more complicated of the two prevailing
present value methodologies. The dividend discount model, or DDM, is the old-
230   •   Stock Value Worksheet



est and best known of present value calculations, first appearing in The Theory
of Investment Value by John Burr Williams in 1938. DDM values a company based
on the view that the asset is worth the present value of the sum of all future divi-
dend payments. To determine the value of an asset based on DDM, and assuming
no change in the dividend rate, divide the annual dividend by the required rate
of return, meaning the rate a rational investor believes can be earned within that
investor’s risk parameters. If you assume the dividend will grow, reduce the
required rate of return by the rate of dividend growth.
      For example, the GLPC shares of Global Pincushion pay a $0.70 dividend;
the company has little likelihood of changing this rate. Assuming a required
return of 6%, the value of the asset is $0.70/.06, or $11.67. If, on the other hand,
Global Pincushion has historically grown its dividend 2% annually, then its value
would be $0.70/(.06 – .02), or $0.70/.04, or $17.50.
      DDM has a satisfying simplicity. And it is widely available, enabling like
comparisons across disparate universes—one of our favorite criteria for leveling
the valuation field. On Bloomberg.com, you can enter any ticker, the Equity key,
and type “DDM” and you are transported instantly to the dividend discount
model for that stock. You can tweak the inputs such as required rate of return of
dividend growth assumption.
      We don’t use DDM in our models. You might assume that’s because not
every company pays a dividend. There are workarounds to that issue. You can
calculate DDM based on a company’s dividend-paying potential, even if that
company pays a dividend in distant years or not at all. Or you might assume we
distrust DDM because dividends are inherently unstable; in the 2008–2009 reces-
sion, dividends were the early ballast jettisoned in a (usually futile) effort to save
a sinking ship.
      Neither consideration would dissuade us from DDM. The reason we don’t
use DDM is because dividend payments are part of a company’s financial struc-
ture decision and have nothing directly to do with operations. And, quite simply,
stocks trade on operating performance, not financial structure.
      For example, there is a small subset of technology companies that have no
debt, still hold a significant portion of their founding cash, but have tiny opera-
tions. Their balance sheets are blue chip, their dividend-paying capacity is signif-
icant—and investors won’t touch them. Many of these issues (uncharitably called
dead men walking) trade at or near cash per-share; during the 2008–2009 reces-
sion, several traded at a discount to cash. That still failed to spark interest among
investors, who invariably seek operating progress in technology names.
      Dividends alternate between academic approbation and disapproval. The
mid-1990s, a time when I was pursuing a CFA and in touch with academic finan-
                         Present Value Modeling and the Stock Value Worksheet   •   231



cial thinking, was a period in which dividends were viewed as somewhere between
unclean and evidence of slothful management. It has been widely noted that
dividends are responsible for up to 40 percent of total market return in the mod-
ern era, and now they have swung back into favor—at least until they swing out
again. Corporate boards can be investment fashion followers. One such swing in
the prevailing attitude can undo a company’s historical dividend pattern—and
with it the most thoughtfully crafted DDM model.
      Dividends should probably be paid based on a company’s operating prog-
ress—but then again, maybe not. Certain categories, such as real estate invest-
ment trusts (REITs) and master limited partnerships (MLPs), vary their payouts
based on operations. But many investors cherish dividend stability because they
require income stability. Aftermarket auto parts distributor and retailer Genuine
Parts Co. (GPC) is one of a handful of companies that have for decades consis-
tently paid a quarterly dividend and raised the payout every year. Amid the col-
lapse in automotive stocks in 2008–2009, the GPS shares of Genuine Parts Co.
held up better than most, largely because its loyal income-oriented investor base
was not about to discard a sure thing in uncertain times.
      Genuine Parts would seem at first blush to have it backward, operating not
to build the enterprise but to steadily pay out its income to its holders. Of course
GPC has it exactly right, for its particular operating model. Perhaps DDM would
be the best way to value Genuine Parts and some other steady dividend plays.
      But companies of this nature are decidedly in the minority. Average com-
panies are obsessed with growth. They’d better be; most investors insist on it, for
better or worse. Given the prevailing valuation drivers, we do not believe the
dividend discount model gives us the best picture of the growing, or failing-to-
grow, enterprise.


“Discount to the Firm and Discount to Equity
The market’s growth and operating obsession will even inform the “flavor” of
discounted free cash flow we choose. New York University distinguished profes-
sor Aswath Damodaran, the reigning authority on DFCF, in his widely read book
Investment Valuation (John Wiley, 1995), distinguishes between “discount to the
firm” and “discount to equity.” The former incorporates capital structure deci-
sions into the modeling process; the latter does not factor in debt. Discount to
the firm is straightforward on a historical basis but requires capital structure
projections on a go-forward basis. The introduction of yet more uncertainty is
among the reasons, but not the principal reason, why we don’t use discount to
the firm.
232   •   Stock Value Worksheet



       The market’s focus on operations does not fully discount implications of
capital structure decisions. Or, to put it another way, the market doesn’t care what
you borrow, if it helps you grow your operations. Whether that’s right or wrong
is immaterial. We have to deal with the weather we get; and in the stock market,
it’s raining growth.


Puts and Takes of Discounted Free Cash Flow Valuation
The core concepts in discounted free cash flow valuation have been around for
some decades but have not always been widely embraced. The wise man from
Omaha (i.e., Warren Buffett) advised investors to invest in things they under-
stand, such as retailers or detergent makers. The same underlying philosophy
of sticking with the familiar informs many of the investment decisions made
by all kinds of investors, from the bulge bracket firms to kitchen-table hedge
funds. While many sophisticated investors like to think they apply all the best
tools to the investment process, at the end of the day they often fall back on,
“What’s the P/E?”
      Many investors ignore DFCF valuation based on its complexity; others cite
the unreliability of the inputs. The second objection is more pertinent. Dis-
counted free cash flow takes a limited subset of inputs and extrapolates them over
an extended period; rather than a leap of faith, it is more of a steady hop farther
and farther away from the current price. Many variables are involved, from asset-
specific data (e.g., betas, growth rates, and ROE) to marketwide inputs (e.g.,
risk-free rate and market risk premium). These data points are used to project
value over an extended period. The market will likely undergo several cycles, and
the asset will undergo many changes, before the projection period is completed.
The underlying exercise—discounting future cash flows back to present value—
is nearly an exercise in futility, in that the likelihood of forecasting the actual cash
flows with a high degree of accuracy is slim.
      And yet the exercise is inherently logical. Historical comparables is a use-
ful look forward—back through the looking glass. But, amid emerging com-
petitors, shifting technologies, new business and service models, and dynamic
market forces, the looking glass of history can quickly turn into a fun-house
mirror. Discounted free cash flow is unequivocally forward looking. Because
cash f lows can be adjusted mid-stream, via two-step and three-step DFCF
modeling, the methodology can encompass assumptions that the enterprise
will undergo changes in response to market forces and amid movement along
its own development path; it can also incorporate new inputs based on your
own changed perceptions.
                         Present Value Modeling and the Stock Value Worksheet   •   233



      No doubt, taking a basket of market and asset inputs and extrapolating
them well into the future introduces variability and risk. Disciplined DFCF mod-
eling incorporates consistent market inputs and inputs for the asset based on
historical and peer trends. Use of a consistent set of rules may not increase the
likelihood of accuracy, but it increases the likelihood that all outcomes will skew
in the same direction.
      Ultimately, DFCF requires a degree of calibration. You cannot take DFCF
out of the box and plug and play. In the real world, modelers gather information
on the variations in projected asset values versus real prices over time; they
learn which assumptions led them astray; and they make the necessary modi-
fications and tweaks. We would argue that you can open a can of historical
comparables and serve it right at the table. Discounted free cash flow models
need to stew for a while, preferably for a few years, before they really begin to
render their full flavor.
      In the preceding chapter we thumped the podium on the risks to DFCF
related to the misleading signals that it sends, given its reliance on growth rates
founded on return on equity in an era when stockholders’ equity is unreliable. In
our next chapter, we blithely explain the concepts underpinning DFCF as a pre-
lude to deploying it. A clear case of dissociative personality disorder? No, and no
again. Discounted free cash flow is the backbone of present value analysis and,
in our view, indispensible in the ever more rapidly involving markets in which
companies play. We simply need to approach DFCF with eyes wide open.
This page intentionally left blank
                                                 Chapter 12
           DISCOUNTED FREE
           CASH FLOW:
           SETTING THE TABLE




Underlying Concepts of Discounted Free Cash
Flow Valuation
The primary concept behind discounted free cash flow is that an asset is worth
the sum of future cash flows and a terminal value. In essence, our task is to deter-
mine base-line cash flows, estimate growth in cash flows from the base line, find
the appropriate discount rate for these cash flows, determine inputs for a termi-
nal value, and create a structure in which we can generate the needed outputs that
sum to the dollar value of the asset. Figure 12.1 shows the completed DFCF valu-
ation model on the Cisco stock value worksheet. This is just a taste, as we have
quite a few concepts to develop and formulas to build before we get there.
       In the pages that follow, we’ll discuss two possible approaches to discounted
free cash flow and show how to construct both or either on the page. Some orga-
nizational aspects are similar in both methods. We also need to make the struc-
ture able to accommodate change, so we don’t have to tear the thing down and
start over every time we want to incorporate new data or push the projection out
a year further. But before we can actually deploy DFCF, we need to understand a
number of key concepts underlying the approach.
       We’ve already laid much of the groundwork for determining company-spe-
cific cash flows in the last set of calculations in the ratios section of our ratios and
                                                                                    235
Figure 12.1

Cisco has seen its beta come down as the once-speculative technology group becomes associated with stable earnings, high cash, and low debt. Within DFCF
modeling, low betas have the effect of reducing discount rates and thus increasing forecast values.

Cisco
Discounted FCF             Calculation                    Dupont ROE                 2004    2005    2006    2007    2008    2009    2010     2011
                                                          Net Income/Sales               23%     23%     22%     22%     20%     17%      17%      18%
Stock Price                        21.05                  Sales/Assets                   62%     73%     66%     65%     67%     53%      53%      60%
Shrs Outstndng ‘09                 5,872                  Assets/Equity                 138%    146%    181%    169%    171%    176%     176%     176%
Mkt Vlu Equity                   123,611                  Return on Equity             19.2%   24.8%   25.9%   24.1%   23.4%   15.9%    16.3%    18.7%
Book Debt ‘09                   10,295.0
Cost of Debt                        7.0%                  Retained Erngs Prcntg         100%     100%    100%    100%    100%    100%     100%     100%
Total Capital                   133,906                   Constant Growth Rate (g)     19.2%    24.8%   25.9%   24.1%   23.4%   15.9%    16.3%    18.7%
Equity/Total Capital              92.3%                   EPS Growth Rate              13.0%    13.0%   13.0%   13.0%   13.0%   13.0%    13.0%    13.0%
Debt/Total Capital                  7.7%                  Avg of 2 Grwth Rates         15.0%    18.9%   15.0%   18.5%   15.0%   14.4%    14.6%    15.8%
Tax Rate ‘09                        20%                   1st 5-Yr LT Growth Rate      12.0%    12.0%   12.0%   12.0%   12.0%   12.0%    12.0%    12.0%
WACC (k)                           9.3%                   2nd 5-Yr LT Growth Rate       9.0%     9.0%    9.0%    9.0%    9.0%    9.0%     9.0%     9.0%
                                                          Terminal Growth Rate          5.0%     5.0%    5.0%    5.0%    5.0%    5.0%     5.0%     5.0%

Risk-Free Rate                      4%                    Beta                           1.20
Market Risk Premium                 5%                    Cost of Equity                9.7%
      Figure 12.1 (continued)


                                                1        2              3         4          5          6        7          8          9       10
      2-Stage Dscntd FCF                    2010     2011            2012      2013       2014       2015     2016       2016       2017     2017    Termnl
      Traditional CF (NI+DDA)             9,992.7 11,205.1
      Free CF (NI+DDA-CX)                 9,042.7 10,205.1
      Free CF - Chng in Trd WC            8,953.3 10,113.5
      FCF per Share                          1.55     1.75            2.01      2.32       2.66       3.06    3.52        4.05       4.66     5.36    115.25
      FCF Pr Shr/                            1.41     1.46            1.53      1.60       1.68       1.76    1.85        1.94       2.03     2.13     45.87
      Sum of DCFs 2002-11         17.40
      Dscntd Terminal Vlu        45.87
      2-Stg DFCF Value           63.27

      3-Stage Dscntd FCF                    2010     2011            2012      2013       2014       2015     2016       2017       2018     2019    Termnl
      Traditional CF (NI+DDA)             9,992.7 11,205.1
      Free CF (NI+DDA-CX)                 9,042.7 10,205.1
      Free CF - Chng in Trd WC            8,953.3 10,113.5
      FCF per Share                          1.55     1.75            2.01      2.32       2.66       2.90    3.17        3.45       3.76     4.10    88.16
      FCF Pr Shr/Discounted                  1.41     1.46            1.53      1.60       1.68       1.67    1.66        1.65       1.64     1.63    35.09
      Sum of DCFs 2002-11        15.94
      Dscntd Terminal Vlu        35.09
      3-Stg DFCF Value           51.03                                                                               Target Price               27
      DCF Avg Value              57.15                                    To Blended Value:         79.4%            To Target Price:       28.3%
      Comparables Stk Vlu        27.63                                    Dvdnd Yld:                 0.0%            Dvdnd Yld:              0.0%
      Peer Calculated Value       9.16                                    Indctd Total Return:      79.4%            Indctd Total Return:   28.3%
      Blended Average            37.77                                    Risk-Adjusted Rtrn:       66.2%            Risk-Adjusted Rtrn:    23.6%
                                                       For HOLD-rated stocks, target price is blended value
237
238   •   Stock Value Worksheet



valuation worksheet. As a refresher, or if you’ve skipped ahead, in the cash flow
portion of our ratios section on the ratios and valuations page we defined some
terms and created some values that we’ll carry forward to the DFCF page. We
defined and calculated basic cash flow as net income plus depreciation and amor-
tization, and basic free cash flow as basic cash flow less capital spending. We
adjusted free cash flow to reflect the changes in cash usage as encompassed by
current assets and liabilities; we further modified this adjustment to include only
the primary movers in working capital, which are inventories and accounts
receivable on the current assets side and accounts payable on the current liabili-
ties side.
       We regard these inputs as (relatively) predictable based on past company
patterns and the economic environment. While this makes for a cooked-down
DFCF model, it excludes less reliable elements. And it makes for a more consistent
calibration when applied uniformly across the coverage universe.


Discount Rates
Given the time value of money, we need to discount cash flows and terminal
value back to present value, but at what rate? We could use a universal rate, such
as “risk-free” coupon rate on a highly secure fixed-income investment, or the
average return for a riskier asset, such as the historical annual stock market
return. Yet every company is different, and each company’s common equity will
be assessed differently in the market. We therefore need to determine a reliable
rate of cash flow growth for each company; that growth will influence the com-
pany-specific rate at which we discount those flows back to present value.
       Having created the mechanism to forecast individual company cash flows
in our ratios and valuations section, let’s now turn to the asset-specific rates
we’ll use to discount those cash flows. To do so, we’re going to define and cal-
culate a few more needed inputs. Bear in mind that these inputs are widely
discussed and exhaustively available in financial texts and on financial Web
sites. Peruse these Web sites and you’ll discover variations and permutations in
the calculation and methodology of inputs. We could immerse ourselves in
minutiae, argue and nitpick, strive to find a common ground—and wander off
the trail. Or, we can use the middle path of widely accepted doctrine in this
area, accept the orthodoxy, and derive some valuations. You already know
which way we’re going.
       We start with example data from the Qualcomm workbook; we’ll assume
we are modeling for the September 2009 year. We’ll examine the concepts before
examining the Qualcomm example.
                                  Discounted Free Cash Flow: Setting the Table •   239



Weighted Average Cost of Capital (WACC): The Concept
Let’s start in the upper-right-hand corner of the worksheet; for shorthand we’ll
call this the WACC (weighted average cost of capital) section. Organizationally,
we’re going to run these inputs down the left-hand side in columns A and B. Your
column B tallies can be adjusted every year; conversely, you can add successive
years in columns C, D, E, and so on. Typically, the column A cell will describe
the input (the stock price in A4), and the column B cells contain the inputs
($43.33 in B4).
      In the A and B columns, we begin with the following descriptions and
inputs:

     • Stock price (in this case, $43.33, imported from the query worksheet);
     • Shares outstanding 2009 (1.655 billion, imported from the annual
       income statement on the R&V worksheet)
     • Market value of equity (stock price times shares)
     • Book value of debt 2009 ($0, imported from the annual balance sheet
       on the R&V worksheet)
     • Cost of debt

      For WACC, we really want the market value of debt. Pricing of fixed-income
securities is a little harder to find; that’s what made Mayor Michael Bloomberg
of New York City a wealthy man. If possible, substitute the market value of debt
for book value; this is particularly important for speculative debt, unkindly called
junk debt. For cost of debt, you can use the average annual interest cost we cal-
culated for our income statement presentation (purists, abide).
      If there is a meaningful premium of book value of debt to market value of
debt, reduce the book value of debt by the discount. For example, assume a com-
pany shows $2.69 billion in book debt; its average interest cost on the income
statement is 6.5%; and its debt trades at $93 against a par of $100. (Bonds have
$1,000 par value but are discussed and traded as though priced at $100.) To adjust
the value of debt, divide market value of debt by book value, producing a value
of .93. Multiply the book debt of $2.69 billion times .93, which comes to $2.501
billion. To adjust cost of debt, divide the book interest cost by .93, which brings
its market cost of debt to 6.989%.
      Now we’ll start to calculate or gather a few more needed inputs:

     • Total capital
     • Equity as a percentage of total capital
240   •   Stock Value Worksheet



      • Debt as a percentage of total capital
      • Tax rate
      • Weighted average cost of capital

      To calculate total capital, add the market value of equity (shares outstand-
ing multiplied by price) plus the market value of debt, as described above. Cal-
culate equity as a percentage of total capital by dividing market value of equity
by total capital. Immediately below that, calculate debt as a percentage of total
capital by dividing market value of debt by total capital. We will also need the tax
rate for our weighted average cost of capital calculation.
      Weighted average cost of capital is the sum of two percentages: cost of debt
and cost of equity. Cost of debt is self-contained for the asset, in that it contains
only asset-specific information. To calculate the cost of equity, however, we need
both asset-specific inputs as well as market inputs.


Cost of Equity
To calculate cost of equity, we use the capital asset pricing model (CAPM). The
CAPM is the asset-specific return calculated as the sum of the risk-free return,
and the market risk premium adjusted by the company’s variability to the mar-
ket, or beta. Again, refer to Figure 12.1. Below the bloc of data in the upper left
corner, we’ve allotted two lines for calculation of cost of equity. In these two lines
are four descriptive as well as four data cells: risk-free rate, beta, market risk
premium, and cost of equity.
      The risk-free rate is the return an investor can expect from an asset with no
risk. The standard risk-free rate is the return on the long bond, meaning the
10-year U.S. Treasury. Here we encounter one of those inputs whose inherent
uncertainty dissuades some investors from adopting DFCF valuation. During the
first half of 2009, the long-bond yield varied significantly—between 2.2% and
4.0% as fears of inflation alternated with ultra-safe-haven investing. In recent
decades, the long yield has mainly ranged in the 3.5% to 6.5% area. So it is per-
haps best to use the long-term trend in this series. Given the more global nature
of fund flows, interest rates will likely remain toward the lower end of the range
in a normal market. In mid-2009, we most commonly used a risk-free rate rang-
ing from 4.0% to 4.5%.
      The market risk premium reflects what investors can expect, on average,
from risky assets such as stocks. The basis for market risk premium is long-term
                                  Discounted Free Cash Flow: Setting the Table •   241



appreciation in a risky metric such as the most common U.S. equity measure, for
instance the S&P 500. You can use a regression series, such as ordinary least
squares, to get smoother picture of appreciation. We exclude return from divi-
dend income; we adjust for inflation; and we use an exponential average, mean-
ing we value the recent inputs more than less recent ones, with a progressive
degradation in the value of older inputs. All of these considerations lead to an
average market risk premium of 4.75% to 5%.
      Beta measures the variation in returns between the asset and a broader mar-
ket measure. U.S. investors are typically most concerned with the relationship
between the individual equity and the S&P 500. In relation to itself, the S&P 500
has a beta of 1.0. Historically, growth companies in industries such as Internet or
biotechnology have had high betas around 1.3 to 1.6; slower growth and regulated
companies, such as electric utilities, have low betas. The most common measure-
ment for beta is five year’s worth of weekly readings; newer companies use a shorter
time period. Hence, if your area of concern includes lots of immature companies,
you’ll want to adjust your measurement period. Betas are widely available, but you
need to take care not to use raw beta if your measurement group is based on
adjusted beta. We use adjusted beta, which adjusts for severe outliers.
      Plenty of models are available for calculating cost of equity using the capital
asset pricing model. All essentially involve a risk-free rate plus beta-adjusted mar-
ket risk premium. Let’s take a spreadsheet where cell B15 shows a risk-free rate of
4.25%; cell B16 shows a market risk premium of 4.7%; and cell J15 shows a beta
of 1.33. The cell holding the CAPM or cost-of-equity formula will look like this:
  B15 ( J15*B16).


Weighted Average Cost of Capital: The Formula
We’re now ready to return to cell B13 to calculate weighted average cost of capital.
The formula has two components: (1) weighted cost of equity and (2) weighted
cost of debt. Weighted cost of equity is the equity percentage of total capital times
cost of equity. Weighted cost of debt is the debt percentage of total capital times
the cost of debt times a tax adjustment. The tax adjustment is equal to 1 minus
the tax rate.
      For our example, let’s put aside Qualcomm for a moment. As of this writing,
the company was considering levering its balance sheet; still, its most current
balance sheet shows no debt. When a company has no debt, its cost of equity and
its weighted average cost of capital are identical.
242   •   Stock Value Worksheet



       Instead, consider a company with equity percentage of total capital of about
78% and debt percentage of total capital of 22%. For this company, the risk-free
rate of 4.25%, market risk premium of 4.7%, and beta resulted in a cost of equity
of 10.1%. For the cost-of-debt calculation, the market cost of debt is 7.0% and the
tax rate is 35%. Our formula looks like this: (B10*J17) (B11*B8*(1 B12)).
It results in a weighted average cost of capital of 8.9%.
       And now the punch line: we are not going to use WACC in the calculation
of discount to equity, which is our prevailing methodology for DFCF. For dis-
count to equity, our primary discount rate will be the cost of equity. An alterna-
tive scheme, discount to the firm, uses WACC as the primary discount rate. We
won’t use discount to the firm because, while it is useful on a historical basis, it
presupposes the modeler will have some accuracy in forecasting capital structure
decisions of the board going forward, and specifically future debt issuance.
       We nevertheless want to have that option available, particularly for finan-
cial firms. Remember our model is a framework on which analysts and investors
can build fuller structures to meet their particular needs. Those needs may
include modeling based on discount to the firm; and for that WACC is a neces-
sary input.
       We are also very interested in how much WACC varies from cost of equity.
Generally (though not always), the higher that debt weighs in the total capitaliza-
tion calculation, the lower the WACC. Figure 12.2 shows Qualcomm’s weighted
average cost of capital and its cost of equity. Because Qualcomm is debt free, the
two figures were an identical 9.7% as of mid-2009.


Forecast Growth
We’ve now calculated the primary discount rates, specific to the asset, that will
be used to discount cash flows back to present value. Next we turn our attention
to the growth metrics we will use.
      There are two primary ways to “get to” the forecast individual cash flows
that will be discounted back to present value. One is to calculate a reliable series
of growth rates for the free cash flows over the discount period out to terminal
value. The other is to model growth of all the principal inputs (earnings, depre-
ciation, capital spending, etc.) over the discount periods and calculate each cash
flow discretely.
      We’ll begin with method 1, in which we model phased growth stages for the
free cash flows over the discount period. This is the more well known and argu-
ably the simpler approach.
                                        Discounted Free Cash Flow: Setting the Table •          243


Figure 12.2

The weighted average cost of capital (WACC) calculation for Qualcomm, done in mid-2009. While
WACC may not be needed in calculating “discount to equity” DFCF, we want it handy in case we transi-
tion the model to “discount to the firm” DFCF.

Qualcomm
Discounted FCF                Calculation

Stock Price                                    43.43
Shrs Outstndng ‘09                           1,655.6
Mkt Vlu Equity                              71,903.5
Book Debt ‘09                                       -
Cost of Debt                                    7.0%
Total Capital                               71,903.5
Equity/Total Capital                         100.0%
Debt/Total Capital                             0.0%
Tax Rate                                        31%
WACC (k)                                       9.7%

Risk-Free Rate                               4.25%       Beta                                  1.15
Market Risk Premium                          4.50%       Cost of Equity                       9.7%




Method One: Forecasting DFCF per Share
Use of DuPont ROE
Our first step is to create an expanded or components-based view of return on
equity, based on the DuPont method. So-called DuPont ROE dates from a finance
executive who early in the twentieth century wanted a more granular look at
return on equity. There are variations on this methodology, which can be
expanded to five steps or more. For our purposes, three-stage DuPont ROE is
sufficient. We want an accurate and detailed read on ROE because it is a key input
in the growth forecast.
      Given that we already calculated one-step ROE on our Ratios and Valua-
tions worksheet, why do we derive a three-stage ROE calculation on our Stock
Value worksheet? First, it acts as a check on the simple one-stage equation we
used on our ratios and valuations worksheet. More important, it provides detailed
information on the source of any expansion or contraction in ROE.
      The three formulas within our DuPont ROE calculation are net margin (net
income divided by sales), asset turnover (sales divided by assets), and equity as a
244   •   Stock Value Worksheet



percentage of assets (assets divided by stockholders’ equity). The disaggregated
or exploded ROE calculation may be of most interest to investors and analysts
monitoring industries or companies sensitive to turnover, net margin, leverage,
or some combination of the three.
      When we look at Corning’s one-stage ROE in Figure 12.3, we see that ROE
is contracting (albeit from and to high levels relative to peers). From 25.6% in
2006, Corning’s ROE slipped to 22.8% for 2007 and to 20.7% for 2008. When we
look at its DuPont ROE, we see that net margin substantially improved and is
thus not the problem. However, revenue generated by each dollar of asset fell
substantially: asset turnover dropped from 40% in 2006 to 31% by 2008. While
that was the principal drag on ROE, we also see that equity as a percentage of
assets fell from 180% in 2006 to 143% in 2008.
      During 2009, Corning’s ROE headed down to the low teens, principally
reflecting effects of the massive global recession. At the same time, and unlike
many rivals, Corning was not engaging in the massive goodwill impairments that
would have artificially boosted ROE. We can see that for this particular downleg
in ROE, the culprit had shifted from revenue on assets to net margin, which
dropped 1,800 basis points between 2008 and 2009.
      Putting aside the recession of 2009, there is a justifiable explanation for the
ROE slippage at Corning between 2006 and 2008. During this period, stockhold-
ers’ equity rose from $7.25 billion in 2006 to $13.44 billion in 2008. Thanks to
the very high margin and fast-growing precision glass business, Corning pulled
off the near miracle of transitioning from a retained deficit of nearly $5 billion
in 2006 to a retained profit of $1.94 billion by 2008. For most companies, declin-
ing ROEs carry no such silver linings. Information on the source of ROE degrada-
tion or growth can steer the modeling process.


Modeling DuPont ROE
On the upper right of the stock value worksheet, we situate the DuPont ROE
calculation. In our example, we’ll reach back to just beyond the end of the five-
year historical period (2004–2008 at the time of this writing) to 2003. We begin
with year 2003 in cell J1. In cell J2, we divide Corning’s 2003 net income by its
sales; both inputs can be linked from the ratios and valuations page. Express this
and all other outcomes in the calculation as a percentage.
      In cell J3, divide sales by assets. Sales is again imported from the income
statement and assets from the balance sheet, both on the ratios and valuations
worksheet. In cell J4, divide assets by equity; both are imported from the ratios
and valuations worksheet. To calculate DuPont ROE, in cell J5 multiply the three
Figure 12.3

DuPont ROE provides an “exploded,” or more detailed, view of the components of return on equity, including net margin, sales divided by assets, and assets divided
by equity.

Corning
                                               DuPont ROE               2003       2004        2005        2006       2007        2008       2009E       2010E
    Discounted FCF        Calculation        Net Income/Sales               -3%        -9%         13%         36%        37%         46%        31%         36%
                                             Sales/Assets                  29%        40%          41%         40%        39%         31%        29%         33%
Stock Price                       15.59      Assets/Equity                197%       255%        199%         180%       160%        143%       144%        145%
Shrs Outstndng ‘09              1,562.5      Return on Equity             -1.6%      -9.1%       10.8%       25.6%      22.8%       20.7%      13.0%        17.2%
Mkt Vlu Equity                 24,359.6
Book Debt ‘09                   1,664.0
Cost of Debt                      6.5%
Total Capital                  26,023.6
Equity/Total Capital             93.6%
Debt/Total Capital                6.4%
Tax Rate                            8%
WACC (k)                          9.4%

Risk-Free Rate                    4.3%       Beta                           1.20
Market Risk Premium               4.5%       Cost of Equity                9.7%
246   •   Stock Value Worksheet



preceding values together. On the illustration, we see that several net loss years
at Corning led to negative ROE through 2004. Beginning in 2005, net margin
turned positive, resulting in a significant ROE swing of nearly 2,000 basis points
into positive territory.
      As a final step in the DuPont ROE process, drag and drop the inputs and
calculations out to the final modeled year. Drag and drop works well because it
is drawn from our annual stacks on the ratios and valuations worksheet.


Constant Growth Rate (g)
Our next set of values is the growth rate, or g. We could simply use historical
earnings growth, based on normalized earnings or actual earnings adjusted with
ordinary least squares regression. But remember that we are modeling cash flows,
not income. Return on equity reflects not only income growth but also asset
efficiency. At the same time, not all income is retained by the entity; some is paid
out in the form of dividends. For our growth rate, we use return on equity mul-
tiplied by the percentage of earnings retained (as opposed to paid out in
dividends).
      We need to import the retained earnings percentage, which we calculated
in the ratios section on our ratios and valuations worksheet. Returning to our
Corning model, we see that the firm paid no dividend in 2003 (it was suspended
in the earlier crisis of 2001–2002). This is expressed in the fact that our imported
retained earnings percentage for that year is 100%. If we drag and drop this
value, we see the dividend reinstated during 2007, to negligible effect; retained
earnings percentage remained in the mid-90s. With a full year of dividends from
2008 on, and with the global recession weighing on earnings, we can see Corn-
ing’s retained earnings percentage drifting lower, to a targeted range in the low
to mid-80s.
      Somewhere there’s a company that grows stably and steadily; it’s most likely
a candy shop on a corner in Duluth, Minnesota. In the real world, the Corning
worksheet shows the immediate challenges in this methodology. In the span of
half a decade, Corning’s growth varied from negative, to the mid-to-upper 20s,
to the low double digits. Based on that all-over-the-map performance, what are
the company’s true growth prospects?
      To get there, we need to introduce some stabilizers. We can use a simple
average of historical g. We can normalize ROE for a set period and use that in our
growth equation. We can season either of those inputs with a qualitative assess-
ment of the company’s prospects over the coming decade. We can rely on the
consensus long-term earnings growth forecast, adjusted for retained earnings
                                  Discounted Free Cash Flow: Setting the Table •   247



percentage. A degree of subjective input is nearly unavoidable. We’ve normalized
earnings across several cycles; any average or normalizing process can also incor-
porate the forecast g values for the modeled years.


Growth Periods
If the number is not too distorted by global crisis or company-specific factors,
constant growth rate, or g, will stand as our forecast growth rate for the first five
years and possibly beyond. Our assumption is that the company’s structure and
strategy—including product suite and development, go-to-market strategy, man-
ufacturing and operating cost structure, financial structure, and other factors—
will enable it to maintain growth at the current pace for a reasonable time.
Five-year projections—for sales, earnings, and other growth rates—are the
financial analysis standard.
      We also need to estimate a growth rate for the following five years. The law
of large numbers tells us that maintaining growth at past rates becomes more
difficult as the enterprise grows in size. New competitive threats will surface;
superseding technologies will emerge; and various known and unknown forces
will coalesce. The logical assumption is that maturity of the enterprise will be
associated with slower growth; this is also an appropriately conservative
approach.
      Finally, for terminal value calculation, we need to project a long-term
growth rate. This growth rate should hold some relationship to national GDP
growth; that relationship should be based on the long-term experience of the
company’s peer group.


Presentation of Growth Periods
The discussion that unfolded in the early pages of this chapter signals the amount
of variability and necessary degree of subjectivity in any entity’s long-term growth
forecast. In our growth presentation, we accommodate multiple inputs as we
attempt to steer toward the best-efforts growth forecast for the company. Figure
12.4 shows the set of growth options we can use for Corning’s DFCF valuation
model. Although we’ll try to use g (or 5-year g, or smoothed g) whenever possi-
ble, we also want some additional growth information on the worksheet.
      We aggregate and present our growth assumptions immediately below the
DuPont ROE and retained earnings percentage presentations. To calculate fore-
cast growth rate, g, multiply DuPont-derived return on equity times retained
earnings percentage. If we first perform this calculation in cell J8 on the Corning
Figure 12.4

This snippet from the Corning stock value worksheet shows the constant growth rate as well as average growth rates, interim period growth rates, and terminal
growth rates that will figure in our DFCF calculation.

Corning
                                       Dupont ROE                     2003        2004       2005     2006    2007    2008    2009E 2010E
Discounted FCF         Calculation     Net Income/Sales                     -3%       -9%         13%     36%     37%     46%      31%  36%
                                       Sales/Assets                        29%       40%          41%     40%     39%     31%      29%  33%
Stock Price                  15.59     Assets/Equity                      197%      255%        199%     180%    160%    143%     144% 145%
Shrs Outstndng ‘09         1,562.5     Return on Equity                   -1.6%     -9.1%       10.8%   25.6%   22.8%   20.7%    13.0% 17.2%
Mkt Vlu Equity            24,359.6
Book Debt ‘09              1,664.0     Retained Erngs Prcntg              100%       100%       100%    100%     93%       87%       80%      84% 5-yr avg.
Cost of Debt                 6.5%      Constant Growth Rate (g)           -1.6%      -9.1%     10.8%   25.6%    21.2%    18.0%     10.4%     14.5%  17.9%
Total Capital             26,023.6     EPS Growth Rate                   15.0%      15.0%      15.0%   15.0%   15.0%     15.0%     15.0%    15.0%
Equity/Total Capital        93.6%      Avg of 2 Grwth Rates                6.7%       2.9%     12.9%   20.3%    18.1%    16.5%     12.7%     14.7%
Debt/Total Capital           6.4%      1st 5-Yr LT Growth Rate           13.0%      13.0%      13.0%   13.0%   12.2%     14.1%     16.1%    16.5%
Tax Rate                       8%      2nd 5-Yr LT Growth Rate           10.0%      10.0%      10.0%   10.0%    10.0%    10.0%     10.0%     10.0%
WACC (k)                     9.4%      Terminal Growth Rate                5.0%       5.0%      5.0%    5.0%     5.0%     5.0%      5.0%      5.0%

Risk-Free Rate               4.3%      Beta                                1.20
Market Risk Premium          4.5%      Cost of Equity                     9.7%
                                  Discounted Free Cash Flow: Setting the Table •   249



worksheet, we can drag and drop to get a sense of likely growth even over the
forecast period of 2009 and 2010.
      In the next line, beginning in cell J9, we collect the consensus or proprietary
forecast long-term (five-year) EPS growth rate for the company. We can use this
as a further stabilizer in our long-term forecast of g, if need be. We allocate the
next line, beginning in cell J10, for an average of our g calculation and the long-
term growth forecast. While we average the two variables in this line, we will only
use this line if average g is too erratic to be predictive. We also show a five-year
average for g in cell R8.
      In the next line, beginning in J11, we position our first five-year long-term
growth rate. In most cases, this will be our calculation of five-year average g
based on DuPont ROE multiplied by retained earnings percentage. Again, if this
input is too erratic or otherwise unusable (a negative number, for instance), we
will incorporate the five-year average from our lines for average of two growth
rate. Note: for easy identification purposes, all five-year averages in this bloc are
bolded; and in row 11, the five-year average begins in 2007 (cell N11). Drag and
drop average five-year long-term growth rate across all periods.
      In the following line, beginning in cell J12, we input our second five-year
growth forecast. This should be based on a variation off the initial five-year
growth forecast based on g. The best gauge for the appropriate degree of variation
will come from the performance of industry rivals. We model the second phase
of growth to be two to four percentage points below the initial growth phase.
Drag and drop across all periods.
      Beginning in cell J13 on our illustration, we model terminal growth. This
is the long-term sustainable growth rate, typically modeled as some variation on
GDP growth and reflective of the industry experience.
      We’ve taken pains to first describe the arbitrary (proprietary) structure of
our stock value worksheet, in Chapter 11, and then to define and explain the key
concepts at work in discounted free cash flow valuation, in Chapter 12. In Chap-
ter 13, our final chapter on the topic, we lay out two conventional means of DFCF
valuation—two-stage and three-stage—and then introduce a slight variation in
which our OLS CAGRs come into play.
This page intentionally left blank
                                               Chapter 13
           DISCOUNTED FREE
           CASH FLOWS:
           TWO METHODS




Calculating DFCF Valuation
Method One, Part One: Two-Stage Discounted Free Cash Flow
The centerpiece of our present value work, determination of asset value based on
DFCF, is also the centerpiece of our stock value worksheet. Here we bring together
the three elements of our process—current period free cash flows, discount rates,
and forecast growth rates—and use them to calculate value of the asset. We will
model and discount free cash flows and terminal value on a two-stage basis, and
then immediately below we will replicate the process on a three-stage basis.
      Figure 13.1 shows Corning’s two-stage discounted free cash flow valuation.
Corning’s two-stage DFCF calculation begins with the assumption that cash
flows will grow faster than earnings. That’s because Corning has invested heavily
in glass tanks for years and now will likely slow its capital spending in a meaning-
ful way.
      Figure 13.2 shows Qualcomm’s two-stage discounted free cash flow valua-
tion. Qualcomm is a cash machine, because of its royalty and licensing business
with unheard-of operating margins of 85% and higher. Even as it invests heavily
in new technologies, its cash flow per share typically exceeds its earnings per
share comfortably. Refer to Figure 13.1 and Figure 13.2 as we go along, to under-
stand the staging of our model.
                                                                                 251
Figure 13.1

Corning’s two-stage DFCF calculation begins with the assumption that cash flows will grow faster than earnings. That’s because Corning has invested heavily in
glass tanks for years and now will likely slow its capital spending in a meaningful way.

Corning
                                               Dupont ROE                  2003      2004       2005     2006         2007      2008   2009E       2010E
Discounted FCF         Calculation       Net Income/Sales                       3%      17%       28%       34%          38%      40%      34%        36%
                                         Sales/Assets                         29%      40%        41%       40%          39%      31%      29%        33%
Stock Price                   15.59      Assets/Equity                       197%     255%       199%      180%         160%     143%     144%       145%
Shrs Outstndng ‘09          1,562.5      Return on Equity                     1.7%    17.3%      23.1%    24.2%        23.2%     17.9%   14.2%       17.4%
Mkt Vlu Equity             24,359.6
Book Debt ‘09               1,664.0      Retained Erngs Prcntg               100%      100%       100%       100%        93%        87%    83%        86% 5-yr avg.
Cost of Debt                  6.5%       Constant Growth Rate (g)             1.7%    17.3%      23.1%      24.2%       21.6%     15.6%   11.7%      14.9%  17.6%
Total Capital              26,023.6      EPS Growth Rate                    15.0%     15.0%      15.0%      15.0%      15.0%      15.0%   15.0%     15.0%
Equity/Total Capital         93.6%       Avg of 2 Grwth Rates                8.4%     16.2%      19.0%      19.6%       18.3%     15.3%   13.4%      14.9%
Debt/Total Capital            6.4%       1st 5-Yr LT Growth Rate            13.0%     13.0%      13.0%      13.0%      16.3%      17.7%   17.1%     16.3%
Tax Rate                        8%       2nd 5-Yr LT Growth Rate            10.0%     10.0%      10.0%      10.0%       10.0%     10.0%   10.0%      10.0%
WACC (k)                      9.6%       Terminal Growth Rate                5.0%      5.0%       5.0%       5.0%        5.0%      5.0%    5.0%       5.0%

Risk-Free Rate               4.25%       Beta                                 1.20
Market Risk Premium          4.70%       Cost of Equity                      9.9%

                                           1           2         3          4          5           6           7            8           9          10
2-Stage Dscntd FCF                       2009        2010       2011       2012       2013        2014        2015         2016        2016       2017       Trmnl
Traditional CF (NI+DDA)                   2,378.6     2,940.3
Free CF (NI+DDA-CX)                       1,479.6     2,042.3
Free CF - Chng in Adj WC                  1,106.6     2,015.7
FCF per Share                                0.71        1.28       1.50      1.76       2.06        2.41           2.83        3.31      3.88       4.54      92.86
FCF Pr Shr/Discounted                        0.64        1.06       1.13      1.21       1.29        1.37           1.46        1.56      1.66       1.77      36.16
Sum of DCFs 2002-11              13.15
Dscntd Terminal Vlu              36.16
2-Stg DFCF Value                49.31
      Figure 13.2

      Qualcomm is a cash machine, thanks to its royalty and licensing business with unusually rich operating margins 85% and higher. Even as it invests heavily in new
      technologies, its cash flow per share typically exceeds its earnings per share comfortably.

      Qualcomm
      Discounted FCF         Calculation           Dupont ROE                         2003      2004       2005     2006         2007       2008   2009E          2010E    AVG
                                                   Net Income/Sales                      33%       34%        37%      33%          37%        28%    15%            28%
      Stock Price                    43.43         Sales/Assets                          45%       47%        45%      49%          48%        46%    46%            56%
      Shrs Outstndng ‘09           1,655.6         Assets/Equity                        119%      112%       113%     113%         117%       145%   137%           130%
      Mkt Vlu Equity              71,903.5         Return on Equity                    17.5%     17.8%      19.0%    18.4%        20.8%      18.8%   9.2%          20.5%
      Book Debt ‘04                       -
      Cost of Debt                    7.0%         Retained Erngs Prcntg                88%       82%        72%        73%         74%       73%          66%      78%
      Total Capital               71,903.5         Constant Growth Rate (g)            15.4%     14.6%      13.8%      13.4%      15.5%      13.8%         6.1%    15.9%   13.5%
      Equity/Total Capital         100.0%          EPS Growth Rate                     15.0%     15.0%      15.0%      15.0%      15.0%      15.0%        15.0%    15.0%
      Debt/Total Capital             0.0%          Avg of 2 Grwth Rates                15.2%     14.8%      14.4%      14.2%      15.2%      14.4%        10.5%    15.4%
      Tax Rate                        31%          1st 5-Yr LT Growth Rate             10.0%     10.0%      10.0%      10.0%      10.0%      10.0%        10.0%    10.0%   10.0%
      WACC (k)                       9.7%          2nd 5-Yr LT Growth Rate              7.5%      7.5%       7.5%       7.5%       7.5%       7.5%         7.5%     7.5%    7.5%
                                                   Terminal Growth Rate                 4.5%      4.5%       4.5%       4.5%       4.5%       4.5%         4.5%     4.5%    4.5%

      Risk-Free Rate                   5%          Beta                                 1.15
      Market Risk Premium              5%          Cost of Equity                      9.7%

                                                     1           2            3        4          5           6            7              8          9            10       10
      2-Stage Dscntd FCF                              2009        2010         2011     2012       2013        2014         2015           2016          2017      2018    Trmnl
      Traditional CF (NI+DDA)                       2,052.1     4,298.4
      Free CF (NI+DDA-CX)                           1,202.1     3,323.4
      Free CF - Chng in Trd WC                      4,180.1     3,297.8
      FCF per Share                                    2.52        1.99        2.26      2.57       2.92        3.31           3.76        4.27          4.85       5.51   106.41
      FCF Pr Shr/Discounted                            2.30        1.66        1.72      1.78       1.84        1.90           1.97        2.04          2.11       2.19    42.26
      Sum of Discreet DFCFs                19.50
      Dscntd Terminal Vlu                  42.26
253




      2-Stg DFCF Value                     61.76
254   •   Stock Value Worksheet



       Beginning at line 19 of Figure 13.2, in column A we set up our headers.
You’ll see lots of familiar concepts. Once again, we will explain the process by
identifying the header in column A and then how to gather or calculate the first
cell input in column C; drag and drop to follow.
       We’ll construct our example as though we were modeling early in 2009
with most of 2009 and all of 2010 in front of us; depending on your own point
in time, adjust all periods. Before we get to the headers beginning in A19, in
line 18 and columns C through L list the single integers 1 through 10; these will
be used in the cash flow discounting process. In cell M18, repeat the integer 10
from cell L18.
       Beginning at A19, we put in the header “2-stage Dscntd FCF” (two-stage
discounted free cash flow); in cells C19 through L19, we put the years 2009
through 2018; and in cell M19, we put the word terminal for terminal value. In
column A, we list headers for the various cash flow items imported from our
ratios and valuation (R&V) worksheet. In cell A20, the header reads “Traditional
CF (NI DDA)” [traditional cash flow (net income plus depreciation, depletion,
and amortization)].
       In column C, we import estimated values for 2009; in column D, we import
estimated values for 2010. In cell C20, we import 2009E cash flow (NI & DDA)
from the R&V page. This value would be approximately in line 35 (depending on
how many historical annuals you’ve modeled, it could be in any column from J
to O). Because we are importing values from our annual stacks on the R&V
worksheet, we can drag and drop the 2010 value into cell D20; but let’s hold off
until we get all the components.


Discreet Period Cash Flows
In cell A21, the header reads “Free CF (NI DDA – CS)” [free cash flow (net
income plus depreciation, depletion, and amortization minus capital spending].
In cell C21, we can get this value by dragging down from cell C20. In cell A22,
the header reads “Free CF Chng in Trd WC” (free cash flow minus change in
trade working capital); in cell C22, import this value from the R&V worksheet.
(Don’t drag and drop from C21, or you’ll get the value for free cash flow minus
the change in traditional, as opposed to trade, working capital.)
      Now that we’ve imported estimated 2009 annual free cash flow, adjusted for
changes in trade working capital, we need to render it on a per-share basis. In cell
A23, the header reads “FCF per Share.” In C23, divide the value in cell C22 by
2009 diluted shares outstanding, imported from the annual income statement
stack on the R&V page. Drag and drop the values from cells C20 through C23
                                      Discounted Free Cash Flows: Two Methods     •   255



into cells D20 through D23. We now have a representation of per-share free cash
flow for the current (2009–2010) period, and this will serve as a baseline for
future growth.
      Our next step is to begin discounting these flows back to present value. The
discount rate we are going to use is the cost of equity. For certain industries, it
may be more appropriate in DFCF valuation to discount at the weighted average
cost of capital; in this case, we discount to the firm, as opposed to discounting to
equity. This requires a few more steps, which is no big deal. I’ve always felt that
the bigger hurdle was the need for reliable insight on a firm’s future capital struc-
ture decisions.
      For the appropriate industries, and assuming you have a good window on
financial structure planning, you can discount to the firm using the weighted
average cost of capital. This will require additional stages in the worksheet to
allow for financial structure assumptions and modeling. For now, let’s continue
with our basic technique that uses free cash flow discounted to cost of equity.
      Even for 2009, we have full-year values that must be discounted back in
time. In cell A24, the header reads “FCF Pr Shr/Discounted” (free cash flow per
share/discounted). Assuming all the structure we’ve put in place, the formula
will read as follows: C23/(1 $J17)^C18. Let’s break it down. C23 refers to the
per-share amount of annual free cash flow. We’re dividing this value by one plus
the asset’s cost of equity, discounted to the appropriate period, which in this case
is one year (cell C18). Drag and drop this value from cell C24 to cell D24. Note
the dollar sign in front of J17, which represents cost of equity; the dollar sign
positioned before J17 enables us to drag this formula across future years without
losing the cost of equity cell value.
      Now that we have a baseline FCF and DFCF per-share values for 2010, we
need to employ a rate of growth to estimate future cash flows. If our g growth
rate is stable and appears reliable, that is the best growth rate to use. If it is highly
erratic, we can stabilize it by averaging it with the consensus or proprietary fore-
cast of long-term (five-year) growth.
      Using the long-term average for g, we calculate estimated 2011 free cash
flow per share in cell E23 with the following formula: D23*(1 $R9). Assessing
our inputs, D23 is the estimated 2010 baseline free cash flow; we are multiplying
this amount by one plus the growth rate in cell R9, which is long-term average g.
We are again using the dollar sign to lock this value for the drag and drop pro-
cess. To discount that 2011 forecast cash flow back to present value, drag and
drop the value from D24 to cell E24. To reiterate, the formula in E24 ( E23/
(1 $J17)^E18) shows that we are discounting the estimated 2011 free cash flow
back two periods or years. Next, drag and drop the values in cells E23 and E24,
256   •   Stock Value Worksheet



for the year 2011, out to cells L23 and L24, representing year 2018. We now have
individual discounted free cash flows that we can sum later.


Terminal Value
The next step is to calculate terminal or ongoing value of the asset. Unlike a bond,
a stock does not mature and discontinue; it has value in perpetuity, as you’d
know if your forebears had been sufficiently prescient to buy scads of GE or IBM
at the initial offering. We operate under the assumption that the companies we
consider for valuation are financially stable and growing at some variation to
normal growth for their industries. Distressed companies or those being valued
for M&A considerations require a different terminal-value calculation.
      By its nature, the value of a perpetuity will be exceptionally sensitive to the
inputs; we’re always careful with inputs, but perhaps extra careful with inputs in
the terminal value calculation. The terminal value of the asset, prior to discount-
ing, is calculated as the last-period projected cash flow divided by a denominator
that incorporates the firm’s long-run required return as well as its growth rate at
maturity. Even for the fastest-growing enterprise, we assume that competitive
forces drawn to an attractive business opportunity as well as the law of large
numbers will combine to bring long-term growth down to some low multiple of
GDP growth.
      For value stocks a well as companies operating in mature industries, fore-
cast terminal growth is 2% to 3%. For growth stocks and companies operating
in nascent or dynamic industries, terminal growth rates of 3% to 5% may be
more appropriate. This latter range puts us in conflict with the DFCF orthodoxy,
which assumes that an asset’s terminal growth rate cannot exceed the economy’s
growth. That might (or might not) be true for a domestic-centered entity; but
with 40% of S&P earnings derived outside the United States, we can fairly ask,
“Which economy?” For long-run required return, we use cost of equity.
      Accordingly, the formula for terminal value of the asset prior to discounting
is L23/(J17 R14), where the numerator L23 represents the final period (2018)
forecast free cash flow and the denominator consists of the cost of equity minus
the forecast long-term growth rate of 4.5%. We then discount this terminal cash
flow value back to final-period present value. To do so, we can drag and drop the
present value discount calculation from cell L24.


Final Value: Two-Stage DFCF Value
Returning to column A, in cell A25 we list the header “Sum of Discreet DFCFs.”
In cell B25, we use the formula, SUM(C24:L24) to sum the DFCF inputs from
                                    Discounted Free Cash Flows: Two Methods   •   257



2009 through 2018. In cell A26 we list the header “Dscntd Terminal Vlu” (dis-
counted terminal value). In cell B26, list this value by importing it from cell M24.
In cell A27, our header reads “2-Stg DFCF Value” (two-stage discounted free cash
flow value). In B27, we sum the values in cells B25 and B26. For Qualcomm, using
these various inputs in midsummer 2009, we arrive at a value of just over $60.


Method One, Part Two: Three-Stage DFCF
Discussion of growth phases within two-stage and three-stage DFCF models begs
the question: why not reduce growth gradually on an annualized basis, for
instance shaving 100 or 75 basis points of forecast growth for each year? The use
of a cumulative projection of several multiyear growth phases achieves that aim
of averaging growth and phasing down to a sustainable rate. Growth will slow.
Invariably and inevitably, success invites competitors who are eager to share in a
prosperous niche. Within the enterprise, earlier risk-taking behavior may be
replaced by a gradual conservatism as management seeks to preserve gains. The
law of large numbers also argues against sustained hypergrowth.
      Use of a three-stage DFCF model is a means of better simulating the normal
growth progression for a company. Typically, firms do not shift from robust
growth to GDP-mimicking growth in a single year. The three-stage model
assumes that after five years at an accelerated rate, a lower average rate will pre-
vail for the ensuing five years as the company works toward its steady-state
growth.
      Three-stage modeling is particularly useful for companies in the midst of an
exceptional growth phase. But we tend to use it for all companies and then weight
its importance based on how well a company meets the three-stage profile.
      From a practical perspective, once all the inputs are in place for two-stage
DFCF modeling, building a three-stage model from the two-stage structure is fast
and simple. While building the three-stage model, review the two-stage DFCF
procedure as outlined previously. Keep in mind that we cannot simply drag and
drop values, but replicating the two-stage structure as the basis for our three-
stage model keeps many of the key formulas intact.


Three-Stage DFCF: Varying the Growth Rate
We’ll build the three-stage model directly beneath the two-stage model; we’ll use
the integers 1 through 10 in line 18 of the two-stage model in our present value
discounting process. Note that Figure 13.3 shows a truncated model in which, to
focus attention, we have excised the two-stage model, which is normally directly
above.
258   •   Stock Value Worksheet



      Figure 13.3 shows three-stage DFCF modeling within our Qualcomm
model. We begin the three-stage model in cell A21 with the header reading “3-Stg
Dscnted FCF” (three-stage discounted free cash flow). Cells C21 through L21
have the years 2009 through 2018. In descending order, we have the header “tra-
ditional cash flow” in cell A22 and that value for 2009 in cell C22. Next is the free
cash flow header in A23 and the 2009 value in C23; continue on down the A and
C columns with free CF change in trade WC, FCF per share, and FCF per
share/discounted. Drag and drop the 2009 values in C22 through C26 to the 2010
column in E22 through E26.
      As a reminder, Figure 13.3 is a truncated view of the normal model in which
two-stage DFCF sits atop three-stage DFCF. We have eliminated the two-stage
portion to better focus on our tasks in the three-stage process. In the full model,
the row values in the formulas will be different than those discussed below.
      As in the two-stage model, we’ll use 2010 estimated free cash flow as a base-
line. We’ll drag and drop the estimated growth formula, D25*(1 $R9) from
D25 (2010) to L25 (2018). And we’ll drag and drop the discounting formula,
  E26/(1 $J17)^E18 from D26 to L26. While in D26, hit F2 and you’ll notice
that the discounting formula references the integers in line 18, used in the two-
stage formula as well. Therefore, if you copy this formula from the two-stage
model, be careful to adjust so that it is linked to the proper discounting cell.
      Now that we have the discreet annual cash flows and their present values in
place, we will adjust them to accommodate another and more temperate phase
of estimated growth. For the first five years, meaning 2009 through 2013, we will
maintain use of the forecast growth rate based on g, or historical average return
on equity adjusted by retained earnings percentage. Practically speaking, because
we’ve hard-modeled our value for 2009 and 2010, that means using the formula,
  D25*(1 $R9) in cell D25 and the dragged-and-dropped variation on it in cells
D25 and G25.
      Beginning in cell H25 for the year 2014, however, we will begin to model a
more temperate pace of growth. Recall that when we set up our estimated growth
grid beneath DuPont ROE, beginning in cell G13 with the header that reads “2nd
5-yr LT Growth Rate,” we modeled a slower growth rate for the five-year period
subsequent to the current five-year forward period. We recommended basing that
growth rate on industry precedents, with inputs from the law of large numbers
and SWOT analysis (strengths, weaknesses, opportunities, threats). The rate of
estimated growth in the second phase, in my experience, ranges between 50%
and 75% of the first growth phase.
      In our Qualcomm model, we’ve assumed that the company will follow its
current 13.5% annualized growth phase with a second phase of 10% annualized
Figure 13.3

Even mighty Qualcomm will slow down over time. We use a five-year forward growth projection based on historical and recent performance, reduced growth
expectations for the ensuing five years, and a much slower growth assumption for the terminal value.

Qualcomm
Discounted FCF         Calculation            Dupont ROE                      2003     2004      2005      2006        2007       2008   2009E         2010E      AVG
                                              Net Income/Sales                   33%      34%       37%       33%         37%        28%    15%           28%
Stock Price                    43.43          Sales/Assets                       45%      47%       45%       49%         48%        46%    46%           56%
Shrs Outstndng ‘09           1,655.6          Assets/Equity                     119%     112%      113%      113%        117%       145%   137%          130%
Mkt Vlu Equity              71,903.5          Return on Equity                 17.5%    17.8%     19.0%     18.4%       20.8%      18.8%   9.2%         20.5%
Book Debt ‘04                       -
Cost of Debt                    7.0%          Retained Erngs Prcntg             88%       82%      72%       73%          74%       73%        66%        78%
Total Capital               71,903.5          Constant Growth Rate (g)         15.4%     14.6%    13.8%     13.4%       15.5%      13.8%       6.1%      15.9%     13.5%
Equity/Total Capital         100.0%           EPS Growth Rate                  15.0%     15.0%    15.0%     15.0%       15.0%      15.0%      15.0%      15.0%
Debt/Total Capital             0.0%           Avg of 2 Grwth Rates             15.2%     14.8%    14.4%     14.2%       15.2%      14.4%      10.5%      15.4%
Tax Rate                        31%           1st 5-Yr LT Growth Rate          10.0%     10.0%    10.0%     10.0%       10.0%      10.0%      10.0%      10.0%     10.0%
WACC (k)                       9.7%           2nd 5-Yr LT Growth Rate           7.5%      7.5%     7.5%      7.5%        7.5%       7.5%       7.5%       7.5%      7.5%
                                              Terminal Growth Rate              4.5%      4.5%     4.5%      4.5%        4.5%       4.5%       4.5%       4.5%      4.5%

                                          1           2           3          4          5         6           7            8            9              10          10
3-Stage Dscntd FCF                      2009        2010         2011       2012       2013      2014        2015         2016         2017           2018       Trmnl
Traditional CF (NI+DDA)                   2,052.1    4,298.4
Free CF (NI+DDA-CX)                       1,202.1    3,323.4
Free CF - Chng in Trd WC                  4,180.1    3,297.8
FCF per Share                                2.52       1.99         2.26      2.57       2.92      3.21        3.53            3.88       4.27          4.70      90.80
FCF Pr Shr/Discounted                        2.30       1.66         1.72      1.78       1.84      1.84        1.85            1.85       1.86          1.87      36.06
Sum of Discreet DFCFs       18.56
260   •   Stock Value Worksheet



growth. That is robust second-phase growth by most company standards; but
Qualcomm is an extraordinary company founded by a true genius whose best
growth markets (i.e., transportation and remote health monitoring) are barely
visible, much less penetrated. Knowledge of the individual company and its
industry—qualitative in addition to quantitative—is essential to the growth
modeling process.
      Accordingly, beginning in cell H25, we adjust the formula, G25*(1 $R9),
where R9 is the 13.5% first-phase growth forecast, to G25*(1 $R12), where R12
is the average second-phase growth forecast of 10%. We don’t change anything in
the terminal value calculations, but note that we now begin with a smaller predis-
counted terminal value than we generated in our two-stage model.


Three-Stage DFCF: Terminal Value and Finishing the Process
The procedure for terminal value for the three-stage model is just as we elabo-
rated in the discussion of two-stage DFCF. To revisit, the formula for terminal
value of the asset prior to discounting is L25/(J17 R14), where the numerator
L25 represents the final period (2018) forecast free cash flow, and the denomina-
tor consists of the cost of equity minus the forecast long-term growth rate of
4.5%. We then discount this terminal cash flow value back to final-period pres-
ent value. In cell M25, the terminal value discounting formula should read
  M25/(1 $J17)^M18.
     Again referencing Figure 13.3, use cell A27 for the header that reads “Sum
of Discreet DFCF,” and in cell B27 sum those discreet cash flows from 2009
through 2018. Use cell A28 for the header “Dscntd Terminal Vlu” (discounted
terminal valuation), and in cell B28 import that value from cell M25. In cell A29,
the header reads “3-stg DFCF Value” (three-stage discounted free cash flow value)
and in cell B29, we sum cells B27 and B28 to get the dollar value of the asset based
on three-stage discounted free cash flow valuation.


Method Two: Discreet Cash Flows from Individual Components
We spent considerable time and energy finding the optimal inputs for current-
period free cash flow valuations, forecast growth rates for future cash flows, and
the appropriate discount factor. Still, some investors may be uncomfortable with
relying on so thin a stream of data—future cash flows—for so much of the valu-
ation calculation. They may also be less enamored of “esoteric” growth rates such
as ROE adjusted by retained earnings percentage. Many investors are more com-
                                     Discounted Free Cash Flows: Two Methods    •   261



fortable with traditional growth analysis yet still want a present value calculation
of future cash flows on their terms.
       For analysts seeking such methodology, we offer an alternative DFCF
method, in which discreet period cash flows are built from their constituents for
all periods, and growth rates are projected based on the historical trends in the
individual inputs. For want of a better name, we call the worksheet “stock value
2”; it could as easily be called “’old-timer’ stock value.” In Figure 13.4, we show
the Qualcomm stock value 2 worksheet; refer to this worksheet throughout our
discussion.
       If we want to forecast individual inputs for the cash flows, we are immediately
met with a challenge: at what rate should we assume growth in the various compo-
nents? One number won’t do for all these different categories. We need a separate
basis for each of these separate inputs. And that basis should encompass various
market conditions, ideally all the various phases of the economic cycle. We have
such a repository of growth rates in our normalized and OLS (ordinary least
squares) worksheet. We’ll import those values for the individual line items.
       If you studied or built two-stage and three-stage DFCF models according to
our formulas, this model will look familiar around the edges (in the upper right
and left sides) where we calculate the growth rates and discount rates. But it looks
somewhat different in the middle, where we perform the actual modeling and
discounting of cash flows.


Setting Up the Stock Value 2 Worksheet
To build this model, we must insert a new worksheet into the workbook. Copy
all of the stock value worksheet, select all on the blank worksheet, and paste that
content onto the new worksheet, which we’ll dub “stock value 2” (if you’ve got a
better name, go for it). We’re going to leave intact everything from line 1 to line
19, including the integers 1 through 10 in line 19 which we’ll use for present value
discounting. We need not re-create the wheel in terms of recalculating discount
rates specific to the asset (at the upper left of the worksheet), and the growth rates
we calculated can stay where they are (at the upper right of the worksheet).
       In the area where we formerly calculated two-stage DFCF, we’ll gather the
components of discounted free cash flow. But first, we need to line up the growth
rates for these components; we’ll do so in the cells formerly devoted to three-
stage discounted free cash flow valuation.
       Beginning at line 33, we gather normalized growth information from our
ordinary least squares (OLS) and normalized worksheet. In cell A33, type in
Figure 13.4

Another present value calculation for modeling purposes uses historical growth rates to grow out all the major components in the free cash flow calculation for all
periods. In Qualcomm’s case, this methodology derived a value of $60 versus a $58 dollar value from the blended two-stage and three-stage traditional method.

Qualcomm
Discounted FCF         Calculation        Dupont ROE                       2003       2004     2005      2006      2007      2008   2009E        2010E
                                          Net Income/Sales                    33%        34%      37%       33%       37%       28%    15%          28%
Stock Price                    43.43      Sales/Assets                        45%        47%      45%       49%       48%       46%    46%          56%
Shrs Outstndng ‘05           1,694.3      Assets/Equity                      119%       112%     113%      113%      117%      145%   137%         130%
Mkt Vlu Equity              73,581.3      Return on Equity                  17.5%      17.8%    19.0%     18.4%     20.8%     18.8%   9.2%        20.5%
Book Debt ‘04                       -
Cost of Debt                    7.0%      Retained Erngs Prcntg               88%       82%       72%       73%        74%      73%       66%       78%
Total Capital               73,581.3      Constant Growth Rate (g)           15.4%     14.6%     13.8%     13.4%     15.5%     13.8%      6.1%     15.9%     13.5%
Equity/Total Capital         100.0%       EPS Growth Rate                    15.0%     15.0%     15.0%     15.0%     15.0%     15.0%     15.0%    115.0%
Debt/Total Capital             0.0%       Avg of 2 Grwth Rates               15.2%     14.8%     14.4%     14.2%     15.2%     14.4%     10.5%     65.4%
Tax Rate                        31%       1st 5-Yr LT Growth Rate            12.0%     12.0%     12.0%     12.0%     12.0%     12.0%     12.0%     12.0%
WACC (k)                       9.7%       2nd 5-Yr LT Growth Rate             8.0%      8.0%      8.0%      8.0%      8.0%      8.0%      8.0%      8.0%
                                          Terminal Growth Rate                4.5%      4.5%      4.5%      4.5%      4.5%      4.5%      4.5%      4.5%      4.5%

Risk-Free Rate                   5%       Beta                                 1.15
Market Risk Premium              5%       Cost of Equity                      9.7%
      Figure 13.4 (continued)


      Qualcomm

                                               1        2           3           4         5           6          7            8          9         10         10
      Component-Based DFCF                      2009     2010        2011        2012      2013        2014       2015         2016       2017       2018     Trmnl
       Net Income                               1,577    3,773       4,573       5,543     6,718       8,143      9,869      12,770      16,524     21,382
       Depreciation                               475      525          587        657       735         823        921        1,030      1,152      1,289
       Capital Spending                        (850)     (975)     (1,288)     (1,702)   (2,249)     (2,972)    (3,926)      (5,188)    (6,855)    (9,057)
       Chng in TWC                               (50)     (50)         (54)       (59)      (64)        (69)       (75)         (82)       (89)       (96)
       FCF                                      1,152    3,273       3,818       4,439      5,141      5,925      6,788        8,531     10,733     13,518
       Shares                                   1,656    1,655        1,716      1,779     1,845       1,914      1,984       2,058       2,134      2,213
       FCF per Share                             0.70     1.98         2.23       2.49       2.79        3.10      3.42          4.15      5.03        6.11   118.04
       FCF per Share Dscntd PV                   0.63     1.64         1.69       1.72       1.76       1.78        1.79        1.98        2.19      2.43     46.88
      Sum of DCFs 2002-11             17.61
      Dscntd Terminal Vlu            46.88
      Component DFCF Value           64.49

      Normalized Growth          Long Term 5-Year            Dscntd Terminal Vlu
      Net Income                        29%     21%
      Depreciation                       12%    32%
      Capital Spending                  32%     56%
      Chng in TWC                         9%    10%
      Shares                              4%      0%


      DCF Avg Value                    64.49                                                                    Target                              56.00     54.94
      Comparables Stk Vlu              58.03                 To Blended Value:                                  To Target:
                                                             Implied Appreciation                   39.6%       Implied Appreciation               28.9%
      Blended Value                    60.62                 Dividend Yield:                        1.04%       Dividend Yield:                    1.04%
                                                             Implied Total Return:                  40.6%       Implied Total Return:              30.0%
263




                                                             Risk-Adjusted Return                   35.3%       Risk-Adjusted Return               26.1%
264   •   Stock Value Worksheet



“Normalized Growth.” In descending order, we then list headers for the following
inputs:

      •   Net income (in cell A34)
      •   Deprecation and amortization (in cell A35)
      •   Capital spending (in cell A36)
      •   Change in trade working capital (TWC) (in cell A37)
      •   Diluted shares outstanding (in cell A38)

      In our OLS calculation, you’ll remember that for every data point we cal-
culated a smoothed growth trend for the longest period permitted by our histori-
cal data compilation. We also maintained a most recent five-year tally of smoothed
trend-line growth. We’ll stay with Qualcomm, which has historical data stretch-
ing back to 1996 and which thus reflects both good and bad times for the econ-
omy and company.
      On the stock value 2 worksheet, we’ll import both smoothed growth rates
both for the longest period permitted by our historical data compilation and for
the most recent five-year period. In cell B33, we have the header that reads “Long
Term”; in cell C33, we put the header “5-year.” We’re now going to import from
the OLS and normalized worksheet:

      • For net income, OLS longest-term growth in cell B34 and 5-year growth
          in cell C34
      • For deprecation and amortization, OLS longest-term growth in cell B35
          and five-year growth in cell C35
      • For capital spending, OLS longest-term growth in cell B36 and five-year
          growth in cell C36
      • For change in trade working capital, OLS longest-term growth in cell
          B37 and five-year growth in cell C37
      • For diluted shares outstanding, OLS longest-term growth in cell B38
          and five-year growth in cell C38

     Qualcomm is an uncommonly prosperous and well-run company, and its
historical data provides an excellent occasion to discuss the objective recording
of hard data and subjective analysis of go-forward prospects. Its long-term net
income growth since 1999, even on smoothed basis, approaches 30%; its five-year
growth is closer to 20%. To hammer home the importance of using OLS, a simple
point-to-endpoint compound annual growth rate (CAGR) calculation signals
growth topping 50 percent!
                                    Discounted Free Cash Flows: Two Methods   •   265



      One reason net income growth has slowed is that Qualcomm is becoming
a bigger and more complex organization, seeking to enter and participate in more
markets. It is no surprise, then, that that pattern of faster long-term growth and
slower five-year growth is reversed for the cash flow items depreciation and
amortization (D&A) and CapEx (i.e., capital expenditures). As the company has
matured, it has purchased assets and grown its spending in support of existing
and new assets. Therefore, five-year trends in D&A and CapEx are much higher
than the long-term trend.
      We have used the actual change in trade working capital in our two-stage
and three-stage DFCF calculation, but we had not previously measured the
smoothed trend in this series. We have inserted a line in our OLS series in which
we first calculate change in trade working capital and then measure the change,
both on an absolute and smoothed basis.
      Interestingly, Qualcomm’s healthy balance sheet illustrates why we use
change in trade working capital rather than absolute working capital. In late
2008, Qualcomm received an enormous one-time cash payment from Nokia as
part of the royalty settlement between those two companies. Had we recorded
change in absolute working capital, Qualcomm would appear to have sunk enor-
mous sums into its working capital and thus would be seen as cash profligate.


Component-Based DFCF Calculation
In the area where we formerly calculated two-stage DFCF based purely on grow-
ing cash flows, we’ll work with an expanded set of inputs. In cells C19 through
L19, we have the integers 1 through 10; repeat the 10 in cell M19. In cell A20 we
have the header reading “Component-Based DFCF”; in cells C20 through L20, we
have the years 2009 through 2018, and in cell M20 we have “Trmnl”
(terminal).”
      Now lets list the components in the A column, along with the forecast inputs
for each in column C (the year 2009) and D (2010). These inputs for the modeled
2009 and 2010 years we will import from our annual stacks on the ratios and
valuations page. In cell A21 is the header “Net Income”; cell C21 is linked to this
value for 2009 and shows the value ’QCOM Rts&Vltns’!O107. Drag and drop
this value to the right for the year 2010. Follow the same procedure for deprecia-
tion and amortization and capital spending, which are linked to the 2009 and
2010 cash flow statement on R&V. For change in trade working capital for 2009
and 2010, we calculate the value based on data in row 7 of the ratios section of
our R&V worksheet. For 2009, we need to adjust because Qualcomm listed the
payment from Nokia at year-end 2008 as a receivable, grossly distorting the swing
266   •   Stock Value Worksheet



in trade working capital. In cell C25 and D25, the sums of these inputs provide
us with adjusted free cash flow for 2009 and 2010.
      We now need to grow out the individual components to calculate period
cash flows for 2011 through 2018. But we won’t be using the growth data com-
piled in the section under DuPont ROE at the top right. Instead, we’ll use the OLS
data gathered below. Beginning with net income for 2011 in cell E21, we use the
formula, D21*(1 $C34), where D21 is our forecast of 2010 net income and C34
is five-year net income growth, based on an ordinary least squares regression.
We’ve again used the dollar sign to lock in the C34 value as we drag and drop to
the right across all periods.
      For 2011 in column E, we can also drag and drop this formula down for
depreciation and amortization, in cell E22, and for CapEx, in cell E23. In this
case, however, we are going to adjust and use the longer-term trends in deprecia-
tion and amortization (cell B35) and in capital spending (cell B36). The five-year
trends in these inputs are through the roof, in the 30% to 50% annual range; we
do not expect Qualcomm to repeat this hyperinvestment phase on a sustainable
basis. Hence, both D&A (an additive to cash flow) and CapEx (a decrement to
cash flow) are forecast to grow at more moderate rates.
      We can also drag down to copy this formula for the change in trade work-
ing capital in cell E24. In cell E25, we sum all inputs for component-derived
free cash flow for 2011. Now we need to render this value on a per-share basis.
We can replicate the formula that captures OLS growth in the share base with
a copy and paste. Once again, analyst discretion and knowledge is required to
best utilize the share growth data. For the past five years, share growth has been
nonexistent as Qualcomm has offset compensation-related share issuance with
repurchases into Treasury. Going forward, we’ll take the more conservative
tack and assume the share base will grow at the 4% long-term rate rather than
the five-year rate of 0%.
      In lines 27 and 28, we perform the same function as in our earlier DFCF
method: rendering free cash flow on a per-share basis (in cell E27), and adjusting
back to present value (in cell E28). To provide an example, we discount to present
value in cell E28 using the formula E27/(1 $J17)^E19, where E27 is adjusted
free cash flow per share; the discount factor is the cost of capital, expressed as 1
plus cell J17; and we are discounting back from year 3, or cell E19. We can now
drag and drop cells E21 through E28, representing year 2011, all the way across
to L21 through L28, representing the year 2018.
      We now have 10 years worth of free cash flows appropriately discounted to
present value. As in our two-stage and three-stage models, these are summed in
the B column. The header in cell A29 reads “Sum of DCFs 2002-11” (sum of dis-
                                     Discounted Free Cash Flows: Two Methods   •   267



creet cash flows 2002 to 2011). The cash flows are summed in cell B 29, where the
formula is SUM(C28:L28).


Terminal Value of Component DFCF
The procedure for terminal value for the individual component model is identical
to the process used in the two-stage and three-stage DFCF discussion. In cell
M27, the formula for terminal value of the asset prior to discounting is L27/
(J17 J14), where the numerator L27 represents the final period (2018) forecast
free cash flow and the denominator consists of the cost of equity minus the fore-
cast long-term growth rate of 4.5%. We then discount this terminal cash flow
value back to final-period present value. To do so, we can drag and drop the pres-
ent value discount calculation from cell L28. In cell M28, the terminal value
discounting formula should read M27/(1 $J17)^M19.
      In cell A30 is the header “Dscntd Terminal Vlu” (discounted terminal
value), and in cell B30 import that value from cell M28. In cell A31, the header
reads “Component DFCF Value”; in cell B31, we sum cells B29 and B30 to get the
dollar value of the asset. For Qualcomm, our component-based present value
calculation using historical growth rates renders a value around $60, as shown in
Figure 13.4. If we check Figure 13.3 and look in cell B32, we see that the average
value calculated from two-stage and three-stage “traditional” DFCF valuation is
$58. While the two values won’t always be this close, we typically see a reasonably
close proximity.


Sensitivity of DFCF Valuation
We recommend playing around with the various inputs to get a sense of just how
sensitive such a far-horizon calculation is to subtle changes. Here are some of our
real-world observations.
      If you use a lower beta, forecast DFCF value of the asset will increase; and
typically, the rate-of-change relation is inversely correlated. In other words, lower
the beta from 1.2 to 1.1, or by 8.3%, and the forecast asset value rises from $58.96
to $64.82, or by 8.2%. If you reduce the risk-free rate assumption or the market
risk premium assumption, the estimated asset value will increase. The reason is
that beta, risk-free rate, and market risk premium all figure in cost of equity (and,
for that matter, in WACC); if you reduce the cost of equity, you reduce the rate at
which you are discounting cash flows to present value. Conversely, and intui-
tively, increasing any of these inputs raises the cost of equity and lowers the esti-
mated asset value.
268   •   Stock Value Worksheet



      In Figure 13.3, a glance at cells B25 and B26 shows that the discounted ter-
minal value is about twice as much as the sum of discreet DFCFs. I’ve seen model-
ers frown at any relationship in which either input is out of balance, prompting
them to “tweak” the model until a 1/1 relationship is restored. In my view, you
need to assess the underlying asset rather than apply one-size-fits-all formulas.
If you had conducted this exercise in 1909 for General Electric, a 2/1 relationship
between terminal value and 10-year discreet cash flows would have woefully
undercounted the ongoing value creation at GE.
      If, back in 1909, you’d conducted the same exercise with General Buggy
Whip, there’s a fair chance your terminal value estimate overstated the ongoing
prospects for the company. Investors and analysts need to understand the nature,
strategy, structure, and philosophy of the entity being modeled; otherwise, any
nonhistorical number is an unfounded guess, as opposed to the more honorable
“guesstimate.”


Annual Updates
Analysts and investors tend to look about two years out. The standard is to model
the income statement on a quarterly basis for the current year as well as the next
year. Sometime after second-quarter reporting is done, meaning about mid-
August, analysts typically start building their following-year quarterly income
statement model; using our prior example, that would entail modeling 2011. As
we indicated on our discussion of the ratios and valuations input, we always want
to incorporate new data, based on a reasonable degree of confidence in its
validity.
       Updating the stock value model to incorporate a new year’s worth of data is
fairly straightforward. But, as our sensitivity discussion above suggests, an errant
data point can significantly skew the outcome; we want to proceed with great
care in redirecting values. We are greatly aided in this endeavor by use of the F2
(i.e., Function 2) key. As experienced Excel hands know, when you are in a for-
mula cell and you hit F2, you immediately see highlighted any cell that contrib-
utes to the formula. Any time you make a change, use the F2 key to certify that
your data points are where you want them to be.
       As a necessary prerequisite to adding another modeled cash flow year to the
stock value worksheet, you must have (1) modeled the full quarterly income state-
ment to include that year and (2) adjusted the ratios and valuations page so the
new annual stack is in place. Assuming you’ve taken these steps, incorporate a
new year in the model proceeding as follows.
                                     Discounted Free Cash Flows: Two Methods   •   269



       Using Figure 13.3 as our example, drag and drop the 2010 modeled free
cash flow information in cells D20 through D24 into cells E20 through E24.
Between column L (in which 2018 free cash flows are modeled and discounted)
and column M (terminal value), insert a column. Drag and drop L25 (FCF per
share) and L26 (FCF per share/discounted) into M25 and M26. We need to move
the discount periods ahead one year as well; so move the integers 1 through 10
currently in cells C18 through L18 into cells D18 through M18. Select all of the
2009 cells (currently in column C) related to two-stage and three-stage DFCF
calculations; this includes cells B20 through B34. Select Delete and check the
prompt Move Cells Left.
      In cell B25 of the two-stage model, which shows the sum of discreet DFCFs,
you will have deleted inclusion of 2009 but may not yet show inclusion of 2019.
Use the F2 key to adjust to include all 10 discreet cash flow periods. For the three-
stage model, do the same in cell B35.
      In the two-stage model, highlight the nondiscounted terminal value in cell
M23; you’ll see that it is referencing the cash flow from 2018 (now in cell K23)
rather than 2019 (cell L23). Adjust terminal value to reference the final discreet
cash flow. For the three-stage model, do the same in line 33.
      In the upper left, we also want to incorporate the most current data for our
WACC calculation conducted in cells B4 through B13. We had been using year
2009 inputs from the 2009 annual stack on the ratios and valuations page. Change
all these to 2010 inputs. Finally, in our DuPont ROE and growth section in the
upper middle to right, drag and drop an additional year, after first making room
by moving over any calculated averages; adjust those averages.
      Your sheet is now updated with the most current hard-modeled data. While
you’re on the sheet, this is a good time to assess all your underlying assumptions
(risk-free rate, market risk premium, etc.) and update any data points that may
have changed (beta, long-term growth assumptions, etc.).


Completing the Stock Value Worksheet
We finish both our individual cash flow and component-based cash flow work-
sheets with the same group of calculations. The main items include estimation
of blended value, which incorporates both our DFCF work as well as the value
derived from historical comparables (we can also incorporate peer derived value,
discussed in a later section). We also compare current price to our blended value
and calculate the implied total return on a risk-adjusted basis. Comparing this
against the market proxy (i.e., estimated average market return), we can set a
270   •   Stock Value Worksheet



rating on the stock, if need be. Finally, also using our risk metric, we can set a
target price on the stock, if need be. Using the Qualcomm two- and three-stage
DFCF stock value worksheet as an example, shown in Figure 13.5, let’s analyze
these processes one at a time.


Blended Value of the Asset
Regardless of the DFCF method (or methods) you’ve chosen to calculate, we now
need to deploy this information in our overall asset value calculation. We’ve also
calculated historical comparable valuations, and the final output from that work
is linked to the stock value worksheet as well. It is time to begin integrating these
approaches in pursuit of an estimate of the dollar value of the asset.
       In cell A39 we have the header that reads “DFCF Avg Value” (discounted
free cash flow average value). In cell B39, we simply average the values from
B27 (two-stage DFCF) and B37 (three-stage DFCF). Use of both in our value
calculation may raise some eyebrows, but we regard use of both as a conser-
vatism mechanism. In cell A40 we have the header reading “Comparables Stk
Vlu” (comparables stock value); and in cell B40, this value is imported from
our R&V worksheet. We left room in cell A41 for the header to read “Peer
Derived Value” (PDV) and in B41 for this value, which we’ll explain later; in
Qualcomm’s case, we felt this value did not provide a useful or credible data
point.
       In cell A42, we have the header that reads “Blended Value,” and in cell B42
we weight the inputs. Of all the shortcuts, compromises, and “bestimates” we’ll
use in our process, this one may invite the most scrutiny, debate, and criticism.
First, to our weighting scheme: in instances where there is no PDV input, we
typically weight our inputs two-thirds DFCF and one-third historical compa-
rables. In cell B42, the formula reads, (B27 B37 B40)/3, where B27 is two-
stage DFCF value, B37 is three-stage DFCF value, and B40 is historical
comparables value. We tend to overweight the more forward-looking metric
(i.e., DFCF) over the more backward-looking mechanism (i.e., comparables)
because the market itself is always looking forward, even as it’s glancing back
over its shoulder.
       Purists on either side of the DFCF-comparables divide may not be happy
with this commingling. But the market is precisely that, a commingling of
investors’ attitudes, theories, styles and beliefs, seasoned with fear and greed.
We can’t know the market’s secret sauce, including the measures of DFCF and
historical comparables poured into the stockpot. But we know for a fact both
are in the soup.
Figure 13.5

The stock value page is not only the place where we conduct DFCF valuation; it is also the place where we aggregate all the different valuation methodologies
(DFCF, historical comparables, Peer Derived Value), set a weighting scheme, and arrive at dollar value of the asset.

Qualcomm
Discounted FCF             Calculation       Dupont ROE                        2003      2004      2005       2006       2007           2008   2009E 2010E               AVG
                                             Net Income/Sales                     33%      34%        37%        33%        37%            28%    15%   28%
Stock Price                        43.43     Sales/Assets                         45%      47%        45%        49%        48%            46%    46%   56%
Shrs Outstndng ‘09               1,655.6     Assets/Equity                       119%     112%       113%       113%       117%           145%   137%  130%
Mkt Vlu Equity                  71,903.5     Return on Equity                   17.5%     17.8%     19.0%      18.4%      20.8%          18.8%   9.2% 20.5%
Book Debt ‘04                           -
Cost of Debt                        7.0%     Retained Erngs Prcntg               88%       82%       72%        73%             74%      73%         66%       78%
Total Capital                   71,903.5     Constant Growth Rate (g)           15.4%     14.6%     13.8%      13.4%          15.5%     13.8%        6.1%     15.9%       13.5%
Equity/Total Capital             100.0%      EPS Growth Rate                    15.0%     15.0%     15.0%      15.0%          15.0%     15.0%       15.0%     15.0%
Debt/Total Capital                 0.0%      Avg of 2 Grwth Rates               15.2%     14.8%     14.4%      14.2%          15.2%     14.4%       10.5%     15.4%
Tax Rate                            31%      1st 5-Yr LT Growth Rate            10.0%     10.0%     10.0%      10.0%          10.0%     10.0%       10.0%     10.0%       10.0%
WACC (k)                           9.7%      2nd 5-Yr LT Growth Rate             7.5%      7.5%      7.5%       7.5%           7.5%      7.5%        7.5%      7.5%        7.5%
                                             Terminal Growth Rate                4.5%      4.5%      4.5%       4.5%           4.5%      4.5%        4.5%      4.5%        4.5%

Risk-Free Rate                       5%      Beta                                 1.15
Market Risk Premium                  5%      Cost of Equity                      9.7%

                                              1           2         3           4         5          6            7               8           9              10           10
2-Stage Dscntd FCF                          2009        2010       2011        2012      2013       2014         2015            2016        2017           2018        Trmnl
Traditional CF (NI+DDA)                      2,052.1     4,298.4
Free CF (NI+DDA-CX)                          1,202.1     3,323.4
Free CF - Chng in Trd WC                     4,180.1     3,297.8
FCF per Share                                   2.52        1.99        2.26      2.57      2.92       3.31            3.76           4.27      4.85           5.51       106.41
FCF Pr Shr/Discounted                           2.30        1.66        1.72      1.78      1.84       1.90            1.97           2.04      2.11           2.19        42.26
                                                                                                                                                                      (continued)
Figure 13.5 (continued)


Qualcomm

                                        1               2           3           4           5           6              7           8           9          10           10
Sum of Discreet DFCFs      19.50
Dscntd Terminal Vlu        42.26
2-Stg DFCF Value           61.76

3-Stage Dscntd FCF                     2009           2010        2011         2012        2013        2014          2015         2016        2017        2018        Trmnl
Traditional CF (NI+DDA)                 2,052.1        4,298.4
Free CF (NI+DDA-CX)                     1,202.1        3,323.4
Free CF - Chng in Trd WC                4,180.1        3,297.8
FCF per Share                              2.52           1.99          2.26        2.57        2.92        3.21           3.53        3.88        4.27        4.70     90.80
FCF Pr Shr/Discounted                      2.30           1.66          1.72        1.78        1.84        1.84           1.85        1.85        1.86        1.87     36.06
Sum of Discreet DFCFs       18.56
Dscntd Terminal Vlu        36.06
3-Stg DFCF Value           54.62

                                                  To Blended Value:                                                Target                                  56.00        54.94
DFCF Avg Value                58.19               Implied Appreciation                     33.9%                   Implied Appreciation                   28.9%
Comparables Stk Vlu           58.03               Dividend Yield:                          1.04%                   Dividend Yield:                        1.04%
                                                  Implied Total Return:                    34.9%                   Implied Total Return:                  30.0%
Blended Value                  58.14              Risk-Adjusted Return                     30.3%                   Risk-Adjusted Return                   26.1%
                                     Discounted Free Cash Flows: Two Methods   •   273



Risk-Adjusted Return to Blended Value
Now that we have a blended value, we need to determine the amount of return
or loss implied by the span between current price and estimated value. In cell F39,
the header reads “Implied Appreciation”; and in cell H30, we measure it with
formula (B42/B4) 1, where B4 is current price (imported from query) and
B42 is blended value. Because dividend yield is part of the total return package,
in cell F40 we have the header reading “Dividend Yield.” In cell G40, we calculate
current annual yield with the formula ’QCOM Rts&Vltns’!L115/’Stock
Value’!B4, which imports the annual dividend from our annual income state-
ment stack on the ratios and valuations page and divides it by the current price.
       Implied total return (header in cell F41, formula in cell G41) is simply the
sum of implied appreciation and dividend yield. Qualcomm began paying a divi-
dend in mid-decade, and its annual yield adds roughly a percentage point to total
return. Next we risk-adjust or normalize the return. The risk-adjusted return
header is in cell F42, and the simple formula H41/J16 is in cell G42. The for-
mula divides forecast total return by beta?
       Why do we risk-adjust the implied total return? We see it as a necessary step
in the recommendation process. We’ve generated all this data to determine the
dollar value of the asset, but also to contribute to the buy, hold, or sell decision.
Every shop that provides equity investment advice seems to have a different for-
mula for the asset decision. We’ve always let the asset’s relationship to the mar-
ket’s forecast behavior drive the decision.
       On average, the market’s long-run annual total return is about 10%. To be
recommended for purchase, an asset should have the potential to outpace that
normal market return; to be recommended for sale, the asset should meaning-
fully lag the forecast market performance. Yet some assets will be more volatile
than others; beta is the accepted measure of variance from the market norm. We
risk-adjust all stocks to remove from the equation any potential return that might
be related to the asset’s variability.
       If our blended value estimate implies a risk-adjusted total return greater
than 10%, the stock warrants a buy or market outperform rating. If the risk-
adjusted total return falls within a band of –10% to 10%, we rate the stock hold
or market perform. And if the stock’s forecast risk-adjusted total return is less
than –10%, we believe the asset should be sold or alternatively rated market-
underperform. [Regarding nomenclature, some investors dislike the implied
fiduciary relationship attached to buy, hold, and sell ratings and prefer market
perform, outperform, or underperform; others like the unambiguous clarity of
buy, hold, and sell.]
274   •   Stock Value Worksheet



Calculating Target Price and Risk-Adjusted Return to Target
Finally, analysts are often charged with providing target prices on their recom-
mended buys; some shops have target prices on all rated assets. This can be a
finger-in-the-air exercise, but there are methods. In our case, we only assign
target prices to buy-rated assets.
      Here’s a method for determining target prices for buy-rated assets. In cell
M38, multiply the current asset price by the average market return (10%, or 1.10)
and multiply that by the beta (1.15). In Qualcomm’s case, that brought us to a
value fractionally under $55, at a time when the stock was trading near $43.50
and our blended value calculation was $58 on the two- and three-stage DFCF
worksheet and about $60 on the stock value 2 (individual component DFCF).
Investors want some stability in their recommendations; this formula would
cause the target to move around every time you refresh the query page and change
the current price.
      So we “freeze” the target price in cell L38, with the target header in J38.
In cells J39 through J42, we paste the headers from F39 to F42. Cell L39 uses
the formula (L38/B4) 1 to measure the distance between current price and
stabilized target. Cell L40 gathers dividend yield, and cell L41 compiles total
return to target. We then risk-adjust total return to our stable target price in
cell L42.
      We typically determine the buy, sell, or hold decision based on risk-adjusted
return to target price. For example, if risk-adjusted return to target for a buy-
rated asset approaches 10% but is still 15% to 18% away from risk-adjusted total
return to blended value, we’d likely take the buy off and put on a hold.
      Just as a reminder, other factors beyond financial modeling may figure in
the stock-rating decision or purchase decision. These will include economic cycle,
company and industry dynamics, relative competitive advantage, and other
factors.
      Any discussion of discounted free cash flow compressed in a few chapters
is necessarily cursory. Nonetheless, we feel we’ve supplied enough structure for
careful modelers to begin their approach to this complex topic. We’ve also sup-
plied sufficient information to value stocks via DFCF using one or more of several
methodologies.
      We’ve now addressed the two primary valuation methodologies—historical
comparables and discounted present value—broadly accepted in the market. But,
except for a brief discussion on relative P/Es, we’ve barely addressed the stock’s
                                   Discounted Free Cash Flows: Two Methods   •   275



value relationship with the broad market. And we have not yet addressed one of
the most important value relationships: that of a stock to its peer group. In the
final major section of our book, we’ll focus on relational valuation, beginning
with the construction of an industry matrix workbook. The culmination of our
work is in a proprietary valuation methodology based on the interaction of his-
torical and forward valuations for a stock within its peer group.
This page intentionally left blank
                                                             PART 4

           RELATIONAL
           VALUATION:
           THE INDUSTRY
           MATRIX WORKBOOK
           AND PEER
           DERIVED VALUE

A Note on Procedure
So far we’ve been meticulous in describing the formula-building process. The
most complex formulas were used in the present value process. Going forward,
we’re going to take more of a conceptual approach and less of a mechanical
approach. Based on your growing knowledge base and our change in tack, we’ll
continue to deemphasize detailed formula descriptions.


Introduction
Now that you’ve built a (fairly complete) individual company equity workbook,
build about 5 or 10 or 20 more, to fill out your coverage. While that may seem a
staggering request, you’ll want to bring every company in your coverage to an
equally granular level of modeling and valuation analysis. And realize that they
do build easier after the first one. The template was built to be replicable. Under-
                                                                                277
278   •   Relational Valuation: The Industry Matrix Workbook and Peer Derived Value



standing the organizational logic of the workbook was half the chore. We’re not
seeking to minimize the work that goes into each workbook, but the task becomes
somewhat lighter with each company.
      Use an existing individual equity workbook as your template, save it into a
dedicated folder for the new company. For each company and workbook, compile
the appropriate historical financial statement data within the new workbook,
maintaining the organizational conventions. Companies present their financial
statements and supplemental data in subtly different ways; be careful with your
links.
      Company equity workbooks fully unlock their value when they interact.
Your wealth of individual company data represents a trove of industry data as
well; it just needs to be organized.


Purpose of the Industry Matrix
For the first time, we’re now going to step away from our individual equity work-
book and move on to the industry matrix workbook.
      The individual equity workbook has provided lots of value so far, including
the preliminary determination of dollar value of the asset. But let’s step back to
review the analytic process. Financial modeling in and of itself is only part of the
analyst’s job. Equally important are an understanding of the company’s business
and the industry in which the company operates. Analysis is both an art and a
science; in addition to compilation of data and objective analysis of the model’s
output, analysts are charged with both objective and subjective assessment of the
company’s prospects both on its own and in relationship to its peer group.
      Creation of an industry matrix enables objective analysis of financial per-
formance and valuation of the company in relation to its peer group. We hope
you’ve used our lessons so far to build multiple models. If not, it is still possible
to import valuation data for use in peer group comparisons. But for maximum
flexibility with the data, we are operating under the assumption that you are
linking to individual equity workbooks that you’ve built according to our
guidance.
      While we’re at it, let’s more fully address the topic of using publicly available
data in the industry matrix, as opposed to linking to individually crafted equity
models. We acknowledge that links to or importation of publicly available data
sometimes makes sense. For example, if you are covering 10 industrial companies
in a specific niche and there are actually more like 20 meaningful players in the
space, it makes sense to incorporate data from outside your core coverage. But,
given the complexity of the process we’re about to describe, outside data can be a
     Relational Valuation: The Industry Matrix Workbook and Peer Derived Value   •   279



rough fit, so process it with care. And you’ll need some ingenuity to adopt imported
public data to the Peer Derived Value (PDV) process we describe later on.
      We can more readily point to multiple advantages for using data linked
directly to our individual workbooks. First off, links to your own models allows
you to precisely tailor your comparative universe, with no compromises. Publicly
available data can be comprehensive. But on the individual models, we have cre-
ated hundreds of data points that can be easily linked.
      As we have exhaustively documented, we live in an era of adjusted earnings,
but not every company uses the same adjustments. Firms tend to pick and choose
among the various non-cash line items to include or exclude, such as FAS 123R
stock option compensation, intangibles amortization, and in-process R&D write-
offs. They similarly take different tacks in adjusting their one-time items such as
restructuring and impairments.
      Many public data sources provide only GAAP results. For those public data
sources providing adjusted earnings, adjustments are not explicit or transparent.
If you seek a common ground in what you exclude or include, you need to control
the exclusions yourself.
      Increasingly important in a global investing environment, the industry
matrix models provides a place where we can convert data from foreign compa-
nies to a common currency; in the parlance, it is the place where we “dollar up”
companies such as Samsung, BMW, and Siemens that play such important roles
in global business. Dollar conversions can make tangible the relationships we’ve
studied intellectually. For example, intellectually we know how small a role U.S.
companies play in the wireless infrastructure business; it is humbling and it “hits
home” to when you see it in dollar terms.
      Finally, a tacit benefit of using your own data is, well, tactile in nature.
Imported data flies by, crossing the retina but perhaps not imprinting on the
brain. If you input data you’ve compiled, it is somehow more familiar. This can
act as a filter for rogue data points. It may be old-fashioned, but we believe you
need to touch the data with your hands to really have it available in your head.


Organization of the Industry Matrix Workbook
Within the industry matrix workbook, we will have a worksheet (or in some cases
several) devoted to the following information:

     • Query
     • Price performance grid
280   •   Relational Valuation: The Industry Matrix Workbook and Peer Derived Value



      •   Industry data compilations
      •   Company comparisons
      •   Segment worksheet
      •   Weightings
      •   Peer Derived Value (peer-adjusted value)

      In brief, here are what each of these provide:

      • Query page. This is a source of real-time prices. Recall that it is a part
          of each individual company workbook.
      •   Price performance grid. This grid allows the analyst to monitor
          absolute stock price and simple or weighted percentage changes for the
          stocks and the composite on a year-to-date basis, relative to the broad
          market, and/or from the point of time at which a ratings change is
          made. The grid can also provide alerts when an asset falls or appreciates
          by a given percentage from a year-opening price, a high or low price, or
          a target price.
      •   Industry data compilations. This information allows the analyst to
          keep all industry data in a single place; it permits easy and useful
          comparisons of single-market or niche trends in revenues, units, market
          share, and so on. It allows for visual checks of the individual asset
          against the simple or weighted mean for any data series.
      •   Company comparisons. These enable the analyst to track and compare
          peer group long-term trends and forecasts in aggregate for the various
          ratios, valuations, growth rates, and margin measures.
      •   Segment worksheet. This compiles data from the company
          comparisons for a handy tear sheet or for client presentations. It is
          differentiated from the company comparisons worksheet because it
          breaks the companies out by end-customer segments, regional focus,
          and other niche characteristics.
      •   Weightings. These enable market cap–weighted analysis of all outputs,
          which in our view provides a more accurate read on market and
          industry trends than simple averages. Weightings are also an essential
          component of Peer Derived Value.
      •   Peer Derived Value (PDV). Also called Peer Adjusted Value, PDV
          enables fully quantified relative valuation within the peer group (as
          opposed to within the broad market) via a broad range of valuation
          metrics.
     Relational Valuation: The Industry Matrix Workbook and Peer Derived Value   •   281



      Those are our worksheets (or, in some cases, groups of worksheets) in brief.
We needn’t go into great detail on the query sheet, except to say that we need a
larger number of prices this time. We broaden the data set to include every com-
pany in the peer group; potential additions to the coverage group; noncovered
companies, if they are meaningful players in the coverage niche; and the major
averages.
      Let’s move on to a more detailed discussion of the contents and purpose of
each worksheet, along with guidance on how to build and populate each one. In
Chapter 14, you will find explanations of the query page, the price performance
grid, and industry data compilation; in Chapter 15, company comparisons, the
segment worksheet, and weightings are presented; and Chapter 16 addresses Peer
Derived Value.
This page intentionally left blank
                                              Chapter 14
           PRICE AND
           PERFORMANCE ANALYSIS




Every analyst has a place where he or she compiles industry data. Most have
several. And having several industry workbooks, in our view, is a problem. We
solve the problem by using a single industry matrix for any one industry.
       Any public company is a complex entity; it may be tightly integrated in its
operations or somewhat ramshackle. Representing all those integrated or dispa-
rate parts and pieces in a single valuation workbook is tough enough; in our
discussion of individual equity analysis, we’ve expressed our concern about
orphaned information and pledged our quest to leave no data point behind.
       Industry analysis is not arithmetically more complex than individual equity
analysis; it is exponentially more complex. Forget the orphaned data point; in the
rush to compile information on covered and noncovered companies, as well as
the trends shaping the industry and its place in the economy, we risk spawning
whole orphanages.
       To lessen the risk, we split the industry matrix into two broad themes. The
first, covered in this chapter, contains price performance as well as quantifiable
industry data outside the realm of valuation. This frees us to focus solely on a
company’s valuation in relationship to its peers in a separate part of the matrix.
Finally, using our Peer Derived Value (PDV) methodology, we’ll quantify that
peer relationship and send that information back to the individual equity
workbook.
                                                                              283
284   •   Relational Valuation: The Industry Matrix Workbook and Peer Derived Value



      Having designated a blank Excel workbook for our industry matrix, our
tasks in this first industry matrix chapter are to prepare and populate the follow-
ing things:

      • A query page, which is a source of real-time prices.
      • A price performance grid, which allows the analyst to monitor prices,
          measure performance, and set alerts
      • One or more industry data compilations, which record in-industry
          trends in revenues, units, market share, and the like.


Query Page and Price Performance Grid
A first step anytime we create a new workbook is to create a query worksheet. The
query page is linked via the Internet to a financial source and provides real-time
pricing for an asset or a basket of assets. Refer to the individual equity workbook
discussion in Part One for more color on this topic.
      The company comparison worksheet we’ve used in communications equip-
ment dates from early in the decade and has data dating back to 1998; with a
current 17 companies in the matrix, it stretches nearly 500 lines down the work-
sheet. We’ll limit our discussion to the communications equipment matrix; com-
munications semiconductors and EMS companies have matrix workbooks of
their own. But we include pricing for these related sectors in this matrix (as well
as on their own matrices) because they are adjacent to the communications
equipment industry and provide useful information in that capacity.


Overview
The financial world is forever measuring and calculating. One set of measure-
ments is analyst performance. Amid all the number crunching and the desire to
get the model right, it is easy to lose sight of the ultimate goal of making money
for the client. Most analyst compensation is directly or indirectly tied to total-
return performance; the higher up you go, the more your very employment
hinges on your performance.
      The price percentage change worksheet, once developed, allows the analyst to
monitor stock price changes on an absolute and relative basis; it also enables the
analyst to calculate simple or weighted percentage price changes for the compo-
                                                Price and Performance Analysis   •   285



nents and the composite on a year-to-date basis and/or from the point of a ratings
change. You can adjust the grid to alert you when an asset falls or appreciates a
given percentage from a year-opening price, a high or low price, or a target price.
      In short, the action sheet is a comprehensive price manager. It is also there
to tap you on the shoulder when it’s time to change a rating or make the asset
buy-sell decision.
      A grid tracking real-time asset price performance is a vital tool in the value
kit. Let’s turn to the example and step through the various parts and pieces within
this model. Remember, we’ve taken off the training wheels so we won’t be coax-
ing you through every formula; there are, however, one or two formulas that bear
explaining.
      Figure 14.1a and b shows absolute and relative price performance along with
alert functions for separate baskets of communications equipment, communica-
tions semiconductor, and EMS companies. At the bottom of the sheet, there is a
less-detailed price tracker for technology sector bellwethers. Finally on the right
is an area to track performance since ratings changes. Refer to the various sec-
tions as they are discussed in the following subheads.


Year-to-Date Percentage Change
As a starting point, dedicate a worksheet to the pricing grid, which we can call
just that (I call this worksheet “action sheet,” for reasons that will be clear further
on). Begin by listing the companies that will be included in the workbook, along
with their tickers. Analyst performance is frequently measured on a calendar
basis. (If this is not your situation, you can adjust accordingly later.)
       Therefore, in line with each company name and ticker, enter the opening
price for the current year (for example, 2009); if you are following along with the
example, this goes in column C. We then put current price, which is linked to the
query sheet, in column D; and percentage change from open in column E.
       Figure 14.1 also shows companies grouped by several categories in the com-
munications technology area. For each category, we show the group simple aver-
age; we also show simple average return for all the stocks on the page. This gives
us an industry benchmark by which to gauge performance and identify leaders
and laggards. This is all standard practice and useful in its way. But ultimately it
is just another batch of numbers and percentages, swimming before our eyes with
all the other data. Our buy- sell triggers help distinguish the activity of individual
tickers.
Figure 14.1a

The price performance grid within the industry matrix workbook allows asset, peer group, and index price-tracking on a year-to-date or specified-period basis and can
provide alerts that assist in asset-timing decisions. Space constraints prevent us from showing the worksheet exactly as it appears in the workbook; it is simply too wide.
Figure 14.1a, focused on the left side of the worksheet, shows year-to-date price change for the coverage universe and for industry bellwethers, as well as various price
alerts that are triggered in the event that a stock rises or falls notably.

Action Sheet
Pricing                                                                                             generates                     generates                    generates
                                                                                                     action if                     action if                    action if
                                                                                       2009          current         2009          current      Current         current
                                          1/1/2009        Current         % chnge    opening        Price less      opening       Price grtr   cycle high      Price less
Company                      Ticker         Price          Price         from open price minus         than        price plus        than      (CH) price         than
                                                                                       10%           RP-10%           20%          RP+20%                       CH-10%
ADC Telecom                  ADCT                  5.47          7.57          38.4%        4.97                           6.56         action         8.88          action
Adtran                       ADTN                14.88          24.83          66.9%      13.53                           17.86         action        24.58
Alcatel-Lucent                ALU                  2.15          4.70         118.6%        1.95                           2.58         action         4.50
Big Band                     BBND                 5.52           4.23         -23.4%        5.02          action           6.62                        4.08
Ciena                        CIEN                  6.70         12.84          91.6%        6.09                           8.04         action        16.33          action
Cisco Systems                CSCO                16.30          24.03          47.4%      14.82                           19.56         action        23.24
Ericsson                      ERIC                 7.81         10.03          28.4%         7.10                          9.37         action        10.66          action
Extreme Networks             EXTR                 2.34           2.71          15.8%        2.13                           2.81                         3.03         action
Corning Inc.                  GLW                 9.53          15.68          64.5%        8.66                          11.44         action         16.17         action
JDS Uniphase                 JDSU                 3.65           6.96          90.7%        3.32                           4.38         action          7.71         action
Juniper                      JNPR                 17.51         27.53          57.2%      15.92                           21.01         action        27.28
Motorola                     MOT                  4.43           8.48          91.4%        4.03                           5.32         action         8.66          action
Nokia                        NOK                 15.60          14.69          -5.8%       14.18                          18.72                       15.74          action
Nortel Networks                NT                 0.26           0.07         -73.1%        0.24          action           0.31                         0.07
Polycomm                     PLCM                13.51          26.85          98.7%      12.28                           16.21         action        26.44
Sycamore Networks            SCMR                 2.69           2.99          11.2%        2.45                           3.23                         3.21         action
Tellabs                      TLAB                  4.12           7.14         73.3%        3.75                           4.94         action          6.95
AVG                                                                           46.6%
      Figure 14.1a (continued)


      Action Sheet
      Pricing                                                                                       generates                 generates                generates
                                                                                                     action if                 action if                action if
                                                                                       2009          current       2009        current      Current     current
                                         1/1/2009         Current         % chnge    opening        Price less    opening     Price grtr cycle high Price less
      Company                  Ticker      Price           Price         from open price minus         than      price plus      than      (CH) price     than
                                                                                       10%           RP-10%         20%        RP+20%                   CH-10%
      Analog Devices            ADI           19.02             27.41          44.1%       17.29                       22.82        action       29.40       action
      Broadcom                 BRCM           16.97             29.54          74.1%      15.43                        20.36        action       30.55       action
      Cavium                   CAVM           10.51             21.94        108.8%         9.55                        12.61       action       21.48
      Qualcomm                 QCOM           35.83             41.70          16.4%      32.57                        43.00                     46.65       action
      Texas Instruments         TXN           15.52             23.64         52.3%        14.11                        18.62       action       25.00       action
      Vishay Intertechnology    VSH            3.42              7.86        129.8%         3.11                         4.10       action        8.49       action
      AVG                                                                     70.9%
      Celestica                  CLS           4.61              9.53        106.7%         4.19                        5.53       action         9.45
      Flextronics                FLEX          2.56              7.49        192.6%         2.33                        3.07       action         7.54        action
      Jabil Circuit               JBL          6.75             14.35        112.6%         6.14                        8.10       action        12.68
      Sanmina-SCI               SANM           2.82               7.15       153.5%         2.56                        3.38       action          7.15
      AVG                                                                    141.4%

       Total Coverage                                  Average return         68.1%

      For this year            S&P 500        903.25          1,071.49        18.6%        821.14
                               NASDAQ        1577.03         2,139.28         35.7%      1433.66
                                 DJIA       8668.39          9,864.94         13.8%      7880.35

      From peak                S&P 500      1565.00           1,071.49       -31.5%
                               NASDAQ       2861.00          2,139.28        -25.2%
                                 DJIA      14164.53          9,864.94        -30.4%
287




                                                                                                                                                          (continued)
Figure 14.1a (continued)


Action Sheet
Pricing                                                                               generates                 generates                 generates
                                                                                       action if                 action if                 action if
                                                                            2009       current       2009        current      Current      current
                                    1/1/2009      Current      % chnge    opening     Price less    opening     Price grtr   cycle high   Price less
Company                    Ticker     Price        Price      from open price minus      than      price plus      than      (CH) price      than
                                                                            10%        RP-10%         20%        RP+20%                    CH-10%
From peak                 S&P 500      1565.00       1,071.49      -31.5%
                          NASDAQ       2861.00      2,139.28       -25.2%
                            DJIA      14164.53      9,864.94       -30.4%

Sub-Sectors
Enterprise                                                       43.8%
Carrier-Centric                                                  44.7%
Components                                                       64.5%

Bellwethers
Microsoft                  MSFT          19.44        25.55       31.4%
Cisco Systems              CSCO          16.30        24.03       47.4%
Nokia                       NOK          15.60        14.69       -5.8%
Motorola                    MOT            4.43        8.48       91.4%
Qualcomm                   QCOM          35.83        41.70       16.4%
Oracle                     ORCL           17.73       20.74       17.0%
Intel                       INTC         14.56        20.17      38.5%
Hewlett-Packard             HPQ          35.94        47.38       31.8%
Dell                        DELL         10.07        15.81       57.0%
Intl. Business Machines      IBM         83.50       125.93      50.8%
Xerox                        XRX           7.97        7.75       -2.8%
Amazon                     AMZN          49.91        95.71       91.8%
                                                                 38.7%
      Figure 14.1b

      In a compacted view, Figure 14.1b show mainly the right side of the price performance grid, which focuses on relative performance (in relation to the market benchmark),
      composite performance on a market-capitalization-weighted basis, and performance of the asset in the span since a ratings change and/or asset decision not necessarily
      made within the performance-assessment period.


      Action Sheet
      Pricing
                                                                                                                                          Rating       Rating     since
                                              1/1/2009            Relative                Market           MrktCp *                       Change       Change rating %
      Company                      Ticker       Price             Prfrmnc                   Cap                 Rtrn          Rating        Date        Price     Chng
      ADC Telecom                  ADCT              5.47              19.8%               731,458,458        280,815,861     HOLD          2/9/2009         3.52    115%
      Adtran                       ADTN            14.88              48.2%             1,559,005,426      1,042,480,107       BUY       12/19/2006        21.60      15%
      Alcatel-Lucent                 ALU             2.15            100.0%           10,620,566,069     12,596,485,338       HOLD         1/31/2007       12.79     -63%
      Big Band                     BBND             5.52             -42.0%                280,541,215      (65,561,262)      HOLD        8/11/2009          4.00      6%
      Ciena                         CIEN            6.70              73.0%              1,175,300,426     1,077,066,361       BUY        12/8/2006        25.56     -50%
      Cisco Systems                CSCO            16.30               28.8%         139,197,984,228     66,012,295,588        BUY        2/27/2009        14.57      65%
      Ericsson                      ERIC             7.81                9.8%          31,985,669,148       9,091,957,171      BUY          8/8/2007       18.97     -47%
      Extreme Networks              EXTR            2.34               -2.8%               241,078,731          38,119,287     BUY        11/5/2008          2.00     36%
      Corning Inc.                  GLW             9.53               45.9%          24,375,908,954      15,730,518,370       BUY        2/17/2009        11.56      36%
      JDS Uniphase                  JDSU            3.65               72.1%            1,548,320,216      1,404,093,127      HOLD        11/3/2008          5.49     27%
      Juniper                       JNPR            17.51              38.6%           14,431,639,310     8,258,425,236        BUY        12/5/2007         31.79    -13%
      Motorola                      MOT             4.43               72.8%          19,464,685,669     17,795,028,659       HOLD        3/22/2007         17.85    -52%
      Nokia                         NOK            15.60             -24.5%            54,462,115,764    (3,176,956,753)       BUY        1/10/2008        34.69     -58%
      Nortel Networks                NT             0.26              -91.7%                35,851,585       (26,199,235)     HOLD
      Polycomm                     PLCM            13.51               80.1%            2,257,073,861      2,228,672,487       BUY        6/18/2007        33.81        -21%
      Sycamore Networks            SCMR             2.69                -7.5%             850,541,084          94,855,883     HOLD         6/1/2004         4.36        -31%
      Tellabs                       TLAB             4.12             54.7%             2,828,690,161      2,073,457,351      HOLD        1/24/2007         9.95        -28%
                                                                     28.0%                                         43.9%
                                                                                                                                                                   (continued)
289
Figure 14.1b (continued)


Action Sheet
Pricing
                                                                                                            Rating      Rating     since
                                     1/1/2009     Relative        Market         MrktCp *                   Change      Change rating %
Company                    Ticker      Price      Prfrmnc            Cap            Rtrn          Rating     Date        Price     Chng
Analog Devices              ADI           19.02       25.5%      7,991,360,787 3,525,106,046       BUY     10/13/2008       23.02      19%
Broadcom                   BRCM           16.97       55.4%     14,648,886,454 10,850,707,291      BUY       2/2/2007       31.83      -7%
Cavium                     CAVM           10.51       90.1%          910,012,423  989,670,980      BUY       4/2/2008        17.06     29%
Qualcomm                   QCOM           35.83       -2.2%     69,318,328,268 11,356,365,809      BUY      2/23/2009       33.33      25%
Texas Instruments           TXN           15.52       33.7%     29,820,062,590 15,601,733,778      BUY      3/24/2009        17.01     39%
Vishay Intertechnology      VSH            3.42      111.2%       1,466,962,129 1,904,477,150      BUY      4/30/2009         5.87     34%
                                                      52.3%                            35.6%
Celestica                    CLS          4.61        88.1%       2,184,275,939 2,331,157,835     HOLD     10/2/2006        11.19    -15%
Flextronics                  FLEX         2.56       174.0%      6,072,395,976 11,694,106,313      BUY      6/1/2000       31.00     -76%
Jabil Circuit                 JBL         6.75        94.0%      3,065,708,251 3,451,760,401       BUY     5/19/2006       34.00     -58%
Sanmina-SCI                 SANM          2.82       134.9%                    0            -     HOLD      8/2/2006        3.38     112%
                                                     122.7%                           154.4%
                                                                                            -

                                                                                196,160,639,181
 Total Coverage                                      49.4%    weighted return           44.4%

For this year              S&P 500       903.25
                           NASDAQ       1577.03
                             DJIA      8668.39

From peak                  S&P 500     1565.00
                           NASDAQ      2861.00
                             DJIA     14164.53
      Figure 14.1b (continued)


      Action Sheet
      Pricing
                                                                                                Rating   Rating     since
                                          1/1/2009     Relative    Market   MrktCp *            Change   Change   rating %
      Company                    Ticker     Price      Prfrmnc      Cap       Rtrn     Rating    Date     Price     Chng
      Sub-Sectors
      Enterprise                                          25.2%
      Carrier-Centric                                      26.1%
      Components                                          45.9%

      Bellwethers
      Microsoft                  MSFT         19.44       12.8%
      Cisco Systems              CSCO         16.30
      Nokia                       NOK         15.60
      Motorola                    MOT           4.43
      Qualcomm                   QCOM         35.83
      Oracle                     ORCL          17.73       -1.6%
      Intel                       INTC        14.56       19.9%
      Hewlett-Packard             HPQ         35.94       13.2%
      Dell                        DELL        10.07       38.4%
      Intl. Business Machines      IBM        83.50       32.2%
      Xerox                        XRX          7.97
      Amazon                     AMZN         49.91
                                                          20.1%
291
292   •   Relational Valuation: The Industry Matrix Workbook and Peer Derived Value



Buy-Sell Triggers
We can kick it up a notch if we add action alerts to the worksheet. We acknowl-
edge the wealth and variety of asset-tracking software available. Here are a few
quick and easy (and no-cost) action triggers you can add to your price perfor-
mance grid.
       By way of background, many investors have codified stop-loss rules. If you
purchase an asset and it loses a given percentage, sell it and don’t look back.
Investors adopt these rules in recognition of human nature. Some investors set
hard stops; Investor’s Business Daily is famous for advocating this approach, rec-
ommending a hard stop at 8%. Others construct relative stops, meaning they sell
the stock when it declines a given percentage (say 4% to 7%) more than the
market.
       Without delving too deeply into investor psychology, the academic litera-
ture supports that view that in aggregate, we are prudent in the realm of gains
but risk takers in the realm of losses. Investors have no trouble taking money off
the table after a stock has risen 12% or 20% or whatever. But when stocks decline,
unwilling to admit our miscalculation we rationalize: I liked it at $60, I must love
it at $45; other investors will see it’s a bargain and come in at $30; and so it con-
tinues until it lands on the pink sheets.
        Analysts have to analyze and report on so many companies and events,
answer and make so many phone calls, and otherwise serve so many constituencies
that important buy-sell signals can slip by unnoticed. Whether or not you believe
in hard or relative stops, it is good to provide visual clues that may prompt action.
       Accordingly, in column F note that we have listed the year-opening price
for each stock minus 10%. Then, in column G, we have set up an alert should any
stock dip below that 10% reduced price relative to opening price. Having prom-
ised to dispense with the formulas going forward, let’s get right to ignoring that
new rule. Under the header “Generates Action If Current Price Less Than RP [i.e.,
recent price] Minus 10%,” we have this formula in each cell: IF(D5<F5,”action”,” “).
In other words, if the terms in the header are met, the cell will reveal the word
action; otherwise the cell will remain blank. One example in row 7, Alcatel-Lu-
cent, had a tough 2008 but was having a whip-crack 2009; with its 119% YTD
gain, it plainly was not sounding the sell action alert at the time of our analysis.
       The technology sector was enjoying a good 2009, with many of the gaudiest
gains coming even as we were writing this section. We did not know how good it
would be when the year began, but we sensed a rebound was on the way. Accord-
ingly, we set our upside bogey a bit higher. Column H shows 2009 opening price
plus 20%; and in column I, our header tells us what the formula will reveal:
                                                Price and Performance Analysis   •   293



“Generates Action If Current Price Grtr [greater] Than RP+20%.” The formula
   IF(D5>H5,”action”,” “) is simply a reworking of the earlier formula. And
again, if the terms laid out in the header are met—if the stock rises more than
20% from the opening level—the word action appears. At the time we were final-
izing the book, only a handful of communications equipment companies had
failed to meet the criteria, and one was in the process of going out of business.
      The technology sector rally has been so robust, particularly in some early-
cycle areas, such as semiconductors and EMS (electronic manufacturing ser-
vices), that we’ve also set up a “rolling” stop. We want to hold the asset until it
attains the full measure of its return; but we’d like to know if some portion of the
gain has begun to erode, perhaps signaling a deeper downturn on the way. Cer-
tainly, momentum investors graph and follow these kinds of trends much more
rigorously. Accordingly, in column J, we show the high price for the cycle; and in
column K, we have our 10% stop alert. We see in Figure 14.1 that several stocks
have slipped off their cycle highs. For the equity analyst, this may signal that it is
time to take off the buy rating; for the portfolio manager, it may be time to
lighten up or eliminate the holding.
      We recognize that there are a host of asset-software solutions that can price-
track and provide a range of stop and rolling-stop alerts and triggers. But we like
that this sheet is situated within our industry matrix, and that all the worksheets
share links to the query page—meaning that at all times, our real-time informa-
tion is aligned.


Relative Performance
Every investor is interested in performance in relation to the benchmark. Adding
relative performance to this price performance grid is simple. In column M of
Figure 14.1b, we list the stock’s relative performance on a percentage basis. There
are a couple of ways we could express this; for example, we could express stock
out-performance or underperformance as a ratio. But what investors want to
know and what they are constantly asking is, how much more or how much less
has it done than the market?
      To answer that question, we need the benchmark year-to-date return on the
page; in this case, the S&P 500 capital appreciation percentage is in cell E40 (Figure
14.1a). To get ADC’s relative performance, for example, simply subtract the market
performance from ADC’s absolute performance. If you want to drag and drop this
formula down the column, remember you need to lock the value of cell E40 by
inserting a dollar sign before the number. (If you were dragging and dropping
across the row, you lock the value with a dollar sign before the letter.) When we get
294   •   Relational Valuation: The Industry Matrix Workbook and Peer Derived Value



to cell M22, we use the same formula to get relative return of the coverage segment;
and in cell M38, we get relative return for the entire coverage universe.


Market Cap–Weighted Return
We’ll discuss market-weighted returns much more extensively a little further
along in this chapter. In fact, we need to get a bit ahead of ourselves in that in
column O, we link to the market capitalizations that are the core element of our
weightings worksheet in the industry matrix. If you are uncomfortable with
jumping ahead or returning to finish this section later, you can gather the market
cap data in column O from the query sheet, or by getting data for shares out-
standing for the individual assets and multiplying by share price.
      Once the individual market caps are in column O, you can determine the
market-weighted return in column P as follows. In the individual cells of column
P, multiply percentage year to date price change times that asset’s market capital-
ization. In cell P22, we determine market cap–weighted return of the group by
dividing the sum of market cap-times-return by the sum of market cap, as
expressed in formula SUM(P5:P21)/SUM(O5:O21). We see that weighted
return lags simple average return, likely because some heavy hitters (Nokia, most
notably) were lagging at the time of this pricing “snapshot.”
      For the telecom equipment niche, we now have three ways of looking at peer
group year-to-date return, all in row 22. We have calculated the absolute simple
average return in cell E22, the relative simple average return in cell M22, and the
market cap–weighted return in cell P22.
      In mid-August 2009, year-to-date market cap–weighted capital appreciation
for this group was about 29%, compared with an absolute simple average return
of 32% and a relative simple average return of 21%.
      By early October, all those metrics had moved up, but the gap between
simple average and weighted average had widened. In other words, as the recovery
rally intensified, tech investors were less obsessed with blue chips and willing to
try on more speculative names. By printing this snapshot every week or so and
examining return relationships, you can provide similar insights.
      Finally, while we study all return metrics, in our opinion the market-
weighted return figure gives the truest picture of how investors in this area have
fared, based on the available shares in which to invest.


Absolute and Relative Return from the Asset Decision
By “asset decision,” we mean the ratings change by the equity analyst, or the deci-
sion to buy or sell the asset by the portfolio manager. This is the acid test of
                                                 Price and Performance Analysis   •   295



performance; it is the source of pride and regret, and examining this data can
leave you squirming. In this case, we’ve set up the grid to show the nature of the
rating change in column R; the change date in column S; price at the change date
in column T; and percentage change since the rating change in column U.
       How do we stack up? Leave it to the fence-sitting hold rating to muddy the
waters. We’d need to dedicate much more time, space, and computing power to
calculate cumulative analyst performance based on all the ratings changes. Like the
competing methodologies and algorithms, we’d need to figure out how to quantify
the value of the hold ratings. Life would be so much simpler in a binary (buy and
sell) world; the fact that it is not explains the multiple methodologies for calculating
cumulative performance. As a reminder, we are not trying to replace or subvert the
plethora of highly useful performance analysis spreadsheets and software available.
Our purpose is simply to use easily input data to keep a running tally on the per-
formance of individual assets from the time of the asset decision.
       To calculate relative return from the time of the asset decision, in column
Y we need to record S&P 500 price on the date of the asset purchase/sale. To
calculate relative price change, we use a formula that includes the asset’s absolute
price change since the asset decision (ratings change or sale/purchase) minus the
percentage change in the S&P 500 from the date of the asset decision to real time.
The formula looks like this: U5 ((D$50/Y5) 1).
       On this, our first worksheet (excluding the ubiquitous query workbook) in
the industry matrix workbook, we also tally the performance of some industry
bellwethers; and we divvy up the communications equipment stocks further to
see how the sector is performing based on the primary customer group served
(carriers, enterprise, or components). Although the action sheet will be one of
our more compact worksheets, it carries lots of practical information. Remember
that the various value triggers are not designed to send you straight to the trading
desk; they’re meant to send you back to reexamine the individual valuation work-
book so you can make the most informed decision.


Replicating the Action Sheet
As the new year nears, prepare next year’s action sheet; price it with the January
1 prices. You’ll want to save the prior year’s actions sheet as a worksheet within
the matrix workbook; it provides handy access to historical price information.
For example, the near doubling in the price of the JDSU shares as of late 2009 is
best seen in the context of its 73% decline, from $12.06 to $3.65, during 2008.
And Qualcomm’ peer group underperformance in 2009 takes on a new color
when we consider its 9% decline in 2008, a year in which the communications-
semiconductor peer group fell 44%.
296   •   Relational Valuation: The Industry Matrix Workbook and Peer Derived Value



Industry Data Compilations
One of the greatest challenges the analyst or finance professional faces is infor-
mation management. Analysis, like intelligence itself, is a function of scope and
focus. Scope—the breadth (really, the flood) of data and information that threat-
ens to overwhelm us—must necessarily be brought into focus—into the linear
stream of text or the calculated value—to provide value to the investor.
      Information comes at us in splatter fashion; it is disorderly, random, and
accretive. We fish in the flood for what’s useful to our analysis. Yes, we can buy
blocs of tailored data, but if that alone were sufficient, we could fold our valuation
tents and go home. We would argue that the very act of managing the scope of
ordered and random data and directing it into the funnel of useful information
informs the asset decision more viscerally and successfully than perusing a cold
stream of purchased data.
      A more prosaic challenge facing the analyst is orphaned data. It is not
uncommon for an analyst to have a dozen or more spreadsheets for a single stock,
some with company-only data and some with industry data. The asset decision
can be contemplative, but just as often it’s opportunistic; at key moments, mar-
shalling data off spreadsheets that don’t speak to one another can be like herding
those proverbial cats. (Have they ever been herded?)
      In our individual data worksheet, we seek to make every data point contrib-
ute to the value decision, we strive to eliminate duplication, and we keep it timely
with real-time pricing. The industry matrix worksheet is the logical place for all
industry data. In many cases, compiled industry data does not contribute directly
to the dollar-value-of-the-asset process. But it aids in the subjective assessment
process, which for the analyst is arguably as important as objective and quantita-
tive valuation processes.
      The various industry information worksheets in our communications
equipment model are concerned mainly with market share and/or technology
share in various niches, such as mobile handsets, smart phones, communications
semiconductors, and wire-line and wireless infrastructure equipment. Most of
the worksheets are structured to show absolute revenues or units, as appropriate,
for multiple quarters on an historical and (sometimes) modeled-forward basis.
      Let’s take a look at the Mobile Handsets worksheet, illustrated in Figure
14.2. Some of this data is publicly available; data with much more supporting
detail can be purchased. This kind of table can also be put together based on
company information. The data is expressed in handset units shipped (not sold)
in any quarter. The worksheet is mainly concerned with the shifting fortunes of
the top five handset vendors, and it is supported by other worksheets detailing
Figure 14.2

The fortunes of the top five handset vendors can experience some nasty reversals. Although summing unit shipments and determining market share in units still
has value, this data seems to be becoming less relevant as high-margin smart phones grow in popularity and “dumb” phones sink to commodity pricing levels.
Global Mobile Handset Units and Market Share
Source: Companies, Gartner, StrategyAnalytics, IDC, Argus Estimates

Units Shipped (in millions)
Company                1Q05        2Q05      3Q05      4Q05       2005     1Q06      2Q06      3Q06      4Q06       2006     1Q07      2Q07       3Q07      4Q07     2007
Nokia                    54.96       61.21     67.23     82.22    265.61     76.09     77.07     88.13    102.79    344.92    92.05     100.80     111.06    133.19   437.10
Motorola                 30.14       34.26     38.63     41.88    144.92     45.52     50.17     51.88     61.03    209.25     47.62     35.55      36.54     39.29   159.00
Samsung                  24.48       25.03     25.86     28.36    103.75     38.08     25.53     30.38     32.01    116.48     32.10     37.43      47.32     44.35   161.20
Sony-Ericsson             9.91       11.90     13.85     16.12     51.77     13.60     15.28     19.40     26.00     73.64     21.77     24.92      26.86     29.85 103.40
LG Electronics           11.46       13.07     13.51     16.88     54.92     14.51     14.40     14.96      17.83    61.99     16.01      19.10     21.85     23.54    80.50
All Others               50.04       48.22     47.67     49.65    155.88     46.25     46.67     46.31     44.89    161.03    49.49      41.31      33.68     59.82   184.30
TOTAL                   180.99      193.69    206.75    235.10    776.87    234.04    229.12    251.06   284.56     967.30   259.04     259.11     277.30    330.05 1,125.50

                       % Market Share
Market Share
Nokia                    30.4%      31.6%     32.5%     35.0%     32.5%     32.5%     33.6%     35.1%     36.1% 35.7%         35.5%     38.9%      40.1%     40.4%      38.8%
Motorola                 16.7%       17.7%    18.7%      17.8%    17.7%     19.4%     21.9%     20.7%     21.4% 21.6%         18.4%     13.7%      13.2%     11.9%      14.1%
Samsung                  13.5%      12.9%     12.5%     12.1%     12.7%     16.3%     11.1%     12.1%     11.2% 12.0%         12.4%     14.4%       17.1%    13.4%      14.3%
Sony-Ericsson             5.5%        6.1%     6.7%       6.9%     6.3%      5.8%      6.7%      7.7%      9.1%   7.6%         8.4%      9.6%       9.7%      9.0%       9.2%
LG Electronics            6.3%        6.7%     6.5%       7.2%     6.7%      6.2%      6.3%      6.0%      6.3%   6.4%         6.2%      7.4%        7.9%     7.1%       7.2%
Others                   27.6%      24.9%     23.1%     21.1%     19.2%     19.8%     20.4%     18.4%     15.8% 16.6%         19.1%     15.9%      12.1%     18.1%      16.4%
TOTAL                   100.0%     100.0%    100.0%    100.0%     95.1%    100.0%    100.0%    100.0%    100.0% 100.0%       100.0%    100.0%     100.0%    100.0%     100.0%

                                                                                                                                                                     (continued)
Figure 14.2 (continued)
Global Mobile Handset Units and Market Share
Source: Companies, Gartner, StrategyAnalytics, IDC, Argus Estimates

                       Year over Year Percentage Change in Units Shipped
Company Y/Y                                                          1Q06      2Q06     3Q06      4Q06      2006    1Q07     2Q07      3Q07      4Q07     2007
Nokia                                                                 38.4%     25.9%    31.1%     25.0%    29.9%    21.0%    30.8%      26.0%    29.6%    26.7%
Motorola                                                              51.0%     46.4%    34.3%     45.7%    44.4%     4.6%   -29.1%    -29.6%    -35.6%   -24.0%
Samsung                                                               55.6%      2.0%     17.5%    12.9%    12.3%   -15.7%    46.6%      55.7%    38.6%    38.4%
Sony-Ericsson                                                         37.3%     28.4%    40.1%     61.3%    42.2%    60.1%     63.1%     38.5%    14.8%    40.4%
LG Electronics                                                        26.6%     10.2%    10.7%      5.7%    12.9%    10.3%    32.6%      46.1%    32.0%    29.9%
Others                                                                 -7.6%    -3.2%    -2.8%     -9.6%     3.3%     7.0%    -11.5%    -27.3%    33.3%    14.5%
TOTAL                                                                 29.3%     18.3%    21.4%     21.0%    24.5%    10.7%     13.1%     10.5%    16.0%    16.4%

                       Quarter over Quarter Percentage Change in Units Shipped
Company Q/Q                      2Q05     3Q05     4Q05             1Q06     2Q06       3Q06      4Q06              1Q07     2Q07      3Q07      4Q07
Nokia                             11.4%     9.8% 22.3%                 -7.5%    1.3%     14.4%     16.6%            -10.5%     9.5%     10.2%     19.9%
Motorola                          13.7% 12.8%        8.4%               8.7%   10.2%      3.4%     17.6%            -22.0%   -25.3%      2.8%      7.5%
Samsung                            2.2%     3.3%     9.7%             34.3% -33.0%       19.0%       5.4%             0.3%    16.6%     26.4%     -6.3%
Sony-Ericsson                     20.2% 16.3% 16.4%                  -15.6% 12.4%        26.9%     34.1%            -16.3%    14.5%      7.8%     11.1%
LG Electronics                    14.0%     3.4% 24.9%               -14.0%    -0.7%      3.9%     19.2%            -10.2%    19.3%     14.4%      7.7%
Others                            -3.6%    -1.2%     4.2%             -6.9%     0.9%     -0.8%      -3.1%            10.3%   -16.5%    -18.5%     77.6%
TOTAL                              7.0%     6.7% 13.7%                -0.5%    -2.1%      9.6%     13.3%             -9.0%     0.0%      7.0%     19.0%
      Figure 14.2 (continued)
      Global Mobile Handset Units and Market Share
      Source: Companies, Gartner, StrategyAnalytics, IDC, Argus Estimates

                           Units Shipped (in millions)
      Company               1Q08      2Q08    3Q08          4Q08     2008       1Q09      2Q09      3Q09 4Q09E 2009E 1Q10E 2Q10E 3Q10E 4Q10E                            2010E
      Nokia                  119.00 121.38       117.80      113.00    471.18     93.20    103.20    108.50 126.90 431.80 107.80 112.83 116.57 135.57                     472.77
      Motorola                27.40     28.10    25.40        19.20   100.10      14.70     14.80     13.60  12.00    55.10  11.22  11.78  12.78  14.70                    50.48
      Samsung                 46.80    45.70      51.80       55.94 200.24        45.92     47.53     49.19  55.10 197.75    47.94  49.61  51.35  57.51                   206.41
      Sony-Ericsson           22.30     27.70    25.70        24.20    99.90      14.50     13.80      14.10 14.60    57.00  14.50  15.01  15.53  17.40                    62.44
      LG Electronics          24.40    26.70     23.00        25.70    99.80      22.60     23.39     24.21   27.11   97.32  24.13  24.98  25.85  28.95                   103.91
      All Others              59.00    58.00      66.12       69.43 252.55        64.22     66.47     68.79  82.55 282.03    69.34  71.77  74.28  83.20                  298.59
      TOTAL                 298.90 307.58 309.82             307.47 1,223.77     255.14    269.19    278.40 318.26 1,120.99 274.93 285.98 296.37 337.33                 1,194.61

      Market Share         % Market Share
      Nokia                  39.8% 39.5%           38.0%     36.8%     38.5%     36.5%     38.3%     39.0%     39.9% 38.5%     39.2%     39.5%     39.3%     40.2%       39.6%
      Motorola                9.2%    9.1%          8.2%      6.2%      8.2%      5.8%       5.5%      4.9%      3.8%  4.9%      4.1%      4.1%      4.3%     4.4%        4.2%
      Samsung                15.7% 14.9%           16.7%     18.2%     16.4%     18.0%      17.7%     17.7%     17.3% 17.6%     17.4%     17.3%     17.3%     17.1%      17.3%
      Sony-Ericsson           7.5%    9.0%          8.3%      7.9%      8.2%      5.7%       5.1%      5.1%      4.6%  5.1%      5.3%      5.2%      5.2%     5.2%        5.2%
      LG Electronics          8.2%    8.7%          7.4%      8.4%      8.2%      8.9%       8.7%      8.7%      8.5%  8.7%      8.8%      8.7%      8.7%     8.6%        8.7%
      Others                 19.7% 18.9%           21.3%     22.6%     20.6%     25.2%     24.7%     24.7%     25.9% 25.2%     25.2%     25.1%     25.1%     24.7%       25.0%
      TOTAL                 100.0% 100.0%         100.0%    100.0%    100.0%    100.0%    100.0%    100.0%    100.0% 100.0%   100.0%    100.0%    100.0%    100.0%      100.0%

                                                                                                                                                                      (continued)
299
Figure 14.2 (continued)
Global Mobile Handset Units and Market Share
Source: Companies, Gartner, StrategyAnalytics, IDC, Argus Estimates

                       Units Shipped (in millions)
                       Year over Year Percentage Change in Units Shipped
Company Y/Y             1Q08      2Q08    3Q08     4Q08 2008E 1Q09             2Q09      3Q09 4Q09E 2009E 1Q10E 2Q10E 3Q10E 4Q10E         2010E
Nokia                     29.3% 20.4%         6.1% -15.2%    7.8% -21.7%       -15.0%      -7.9% 12.3% -8.4% 15.7%     9.3%  7.4%  6.8%     9.5%
Motorola                 -42.5% -21.0% -30.5% -51.1% -37.0% -46.4%              -47.3%   -46.5% -37.5% -45.0% -23.7% -20.4% -6.0% 22.5%    -8.4%
Samsung                   45.8% 22.1%         9.5% 26.1% 24.2%        -1.9%       4.0%     -5.0%  -1.5% -1.2%   4.4%   4.4%  4.4%  4.4%     4.4%
Sony-Ericsson              2.4%    11.2%     -4.3% -18.9% -3.4% -35.0%         -50.2%    -45.1% -39.7% -42.9%   0.0%   8.7% 10.2% 19.2%     9.5%
LG Electronics            52.4% 39.8%         5.3%   9.2% 24.0%        -7.4%   -12.4%       5.3%   5.5% -2.5%   6.8%   6.8%  6.8%  6.8%     6.8%
Others                    19.2% 40.4% 96.3%         16.1% 37.0%         8.8%     14.6%      4.0% 18.9% 11.7%    8.0%   8.0%  8.0%  0.8%     5.9%
TOTAL                     15.4% 18.7%        11.7%  -6.8%    8.7% -14.6%       -12.5%     -10.1%   3.5% -8.4%   7.8%   6.2%  6.5%  6.0%     6.6%

                       Quarter over Quarter Percentage Change in Units Shipped
Company Q/Q             1Q08     2Q08     3Q08     4Q08             1Q09     2Q09        3Q09 4Q09E          1Q10E 2Q10E 3Q10E 4Q10E
Nokia                   -10.7%     2.0%    -2.9%    -4.1%             -17.5%   10.7%       5.1%  17.0%        -15.1% 4.7%  3.3% 16.3%
Motorola                -30.3%     2.6%    -9.6% -24.4%              -23.4%     0.7%      -8.1% -11.8%         -6.5% 5.0%  8.5% 15.0%
Samsung                    5.5%   -2.4% 13.3%        8.0%             -17.9%    3.5%       3.5% 12.0%         -13.0% 3.5%  3.5% 12.0%
Sony-Ericsson           -25.3% 24.2%        -7.2%   -5.8%            -40.1%    -4.8%       2.2%   3.5%         -0.7% 3.5%  3.5% 12.0%
LG Electronics             3.7%    9.4% -13.9%      11.7%            -12.1%     3.5%       3.5% 12.0%         -11.0% 3.5%  3.5% 12.0%
Others                    -1.4%   -1.7% 14.0%        5.0%              -7.5%    3.5%       3.5% 20.0%         -16.0% 3.5%  3.5% 12.0%
TOTAL                    -9.4%     2.9%      0.7%   -0.8%             -17.0%    5.5%       3.4%  14.3%        -13.6% 4.0%  3.6% 13.8%
                                               Price and Performance Analysis   •   301



handsets by technology (WCDMA, GSM, CDMA2000, etc.) and smart-phone
market share.
      Investors in this industry are chiefly concerned about market share, even
though that is arguably becoming an outdated concept in the era of the smart
phone. The worksheet also shows percentage changes in units on a year-over-year
and quarter-over-quarter basis. This kind of data compilation enables the analyst
to discern trends that won’t be visible within the individual company
workbook.
      One simple benefit of the industry data compilations is that it allows apples-
to-apples comparisons for companies sited in different nations. While the giants
in the information processing industry are mainly U.S.-based (e.g., IBM and
Hewlett-Packard), the giants in communications equipment are frequently from
overseas (e.g., Ericsson, Nokia, and Alcatel-Lucent). One of our industry compi-
lation worksheets is focused on wireless infrastructure sales, and it is denomi-
nated in dollars. Only by translating the tallies in dollars and stacking them up
can we see the dominance of the European companies in this space—and how
fast risers from Asia such as Huawei threaten that dominance.
      As you get to know your coverage industry better, you’ll determine which
industry data points are integral to your analysis. The price performance work
with which we led off this chapter is essential to every industry, of course. In the
following chapter, we turn our attention back to valuation—in this case, the valu-
ation relationship between the individual equity and its peers.
This page intentionally left blank
                                               Chapter 15
           SIMPLE AVERAGE AND
           MARKET-WEIGHTED
           COMPARISONS




The Company Comparison Worksheet
Overview
The basic premise of the company comparison worksheet is that the linking of
substantially every ratio, valuation, margin, and growth rate we’ve calculated in
the individual worksheets to this industry matrix provides both useful and easy
comparison. Later, we’ll use a subset of these individual and common values to
derive dollar value of the equity based on its current variation from its historical
relationship to the peer group.
      Nothing so taxing is on this worksheet. Everything we’ve done so far has
rewarded focus and close attention. The company comparison worksheet is big
and long but not very complex; putting it together is, frankly, tedious. Console
yourself with the fact that the output from this worksheet is useful; stream your
favorite radio station (on low) while building it. The sheet, once finished, turns
lots of raw data into lots of highly informative conclusions.
      This exercise is a bit like steering a big combine harvester over Midwestern
wheat fields: the process is fairly straightforward, but if you lose your concentra-
tion you might slam into the barn. The chief challenge is marshalling lots of data
onto one workbook from many sources. With earlier versions of Excel, and given
constraints on hardware power, we were typically limited to opening a few work-
                                                                                303
304   •   Relational Valuation: The Industry Matrix Workbook and Peer Derived Value



books at a time. Impressive advances in Excel along with massively enhanced
computing power allow you open dozens of workbooks at once. We think keep-
ing just a few open at a time remains good practice.
       We will also caution that in some older versions of Excel, not every change
made on the individual worksheet is captured by the industry matrix if the matrix
is closed while working on the individual workbook. Again, that seems like less
of a problem with recent iterations of the software. Still, as a precaution, any time
you are inserting or deleting rows or columns in linked workbooks, it makes
sense to have open all workbooks that share links.


Setting up the Grid: Relations to Forecast Values
At the very top of the worksheet, we’ve compiled the most important set of con-
clusions: the relationships of our peer group to our calculated value. This is
shown in Figure 15.1. In columns A and B, we begin by compiling the coverage
universe and respective tickers; as a labor saver, you can link to the company
price performance grid for this data. In column C we have current price, linked
directly to the query page or indirectly from the price performance grid.
      In column D, you can see that we have our calculated discounted free
cash f low (DFCF) value for every company; in Column F are the values for
every company based on historical comparables. In Column H are the values
for Peer Derived Value (PDV) (not yet explained; see Chapter 16). And in
Column J are blended stock values, incorporating a weighted mix of DFCF,
comparables, and PDV.
      These calculated values are all available on the individual worksheets. They
are of interest here mainly in how they vary from the current price both individu-
ally and collectively. In columns E, G, I, and K, we see the variation between
current price and estimated value calculated value by, respectively, DFCF, com-
parables, PDV, and blended.
      Discounted free cash flow is the most forward-looking—and, I’d argue,
inherently the most optimistic. Accordingly, the calculated DFCF values for this
group are on average 68% higher than the current price (or were at the time of
this exercise). Equity values determined by historical comparable values are on
average 25% higher than current prices; this sober-sided group is grounded in
the (five-year) past and looks scarcely two years forward. Value determined by
PDV is the closest to current values at just an 18% average premium; this is an
imperfect closed loop, and so its values should be closest to current prices. Finally,
the average premium to blended stock value is 67%, even though the individual
components are very different.
      Figure 15.1

      The information atop the comparison worksheet draws asset value data from all the individual equity workbooks. The significant premium of blended value over
      current price—an average of 63% —is to be expected, given the long-tailed nature of blended value. Also contributing to this significant premium is the horrific
      period (fall 2008 through spring 2009) from which the market and this group were still recovering at the time this chart was prepared.

      Comparison Worksheet
      Communications Equipment
                                                                                                                  Peer
                                                          DFCF          % to                        % to         Derived        % to         Blended        % to
                                            Price         Value         DFCF        Cmprbls         DFCF          Value         PDV         Stk Value       BSV
      ADC Telecom             ADCT               8.65         12.87         49%           7.38          -15%          14.13         63%            9.62           11%
      Adtran                  ADTN              22.66         45.19         99%         26.76            18%          23.95           6%          36.47           61%
      Alcatel                  ALA               3.37         11.36        237%           1.33         -60%            2.52        -25%            8.02         138%
      Big Band                BBND               4.04           7.42        84%           6.21           54%           4.57          13%            6.41          59%
      Ciena                   CIEN              12.29         19.85         62%          13.12            7%           6.63        -46%           25.16         105%
      Cisco Systems           CSCO              21.04          57.15       172%         27.63            31%          25.78         23%           37.77           80%
      Ericsson                 ERIC              9.30         20.26        118%         13.36            44%          10.66          15%          18.47           99%
      Extreme Networks        EXTR               2.57           1.23       -52%           6.25         143%            6.63        158%            6.25         143%
      Corning Inc.             GLW              15.79         36.11        129%         30.54            93%          19.75         25%            41.11        160%
      JDS Uniphase            JDSU               5.85           2.17       -63%         13.90          138%            4.40        -25%            9.40           61%
      Juniper Networks        JNPR              24.47         34.52          41%        28.75            17%          23.20          -5%          32.49           33%
      Motorola                MOT                7.03          6.67          -5%          8.50           21%          10.62          51%           6.25          -11%
      Nokia                   NOK               12.95         33.05        155%         22.04            70%          14.74          14%          26.80         107%
      Polycomm                PLCM              22.71         44.02         94%         33.58            48%          26.93          19%          38.04           67%
      Sycamore Networks       SCMR               3.08           1.66       -46%           1.15         -63%             1.47       -52%             1.52         -51%
      Tellabs                 TLAB               6.59          8.38         27%           7.26           10%           6.41          -3%            7.25          10%
      Average                                                               69%                         35%                         14%                          67%
305
306   •   Relational Valuation: The Industry Matrix Workbook and Peer Derived Value



Driving the Combine: Marshalling Data
We have calculated many data points on the individual worksheets, and we want
to link most, though not all, to the industry matrix. We find it helpful to organize
the data by a few broad categories. These are the categories:

      •   Earnings and revenues
      •   Balance sheet metrics
      •   Margins
      •   Return measures
      •   Valuations

      We’ll break these down further as we go along. The industry matrix work-
books we built over the years have tended to grow up like a city; the neighbor-
hoods predate the city planners. You can take advantage of what we’ve learned in
our hodgepodge construction to make more rational worksheets.
      The guidelines we use when linking earnings from our worksheets apply to
most of the other sections that follows on this worksheet. For that reason, we’ve
included an illustration, even though the literally hundreds of links in this sec-
tion are all invisible on the page. Even if you place emphasis elsewhere, you should
review our work in this section.


Earnings
It is no surprise that earnings have pride of place in the company comparison
worksheet; earnings are where valuation discussions begin and end. Revenues are
nearly as important, as so much operating leverage is tied to the top line.
       Copy the company names and tickers from the relations to forecast value
section at the top of the worksheet. You are going to want to link to the historical
annual earnings for as many years as you have. It is likely you have a wide range
of historical annual earnings for the companies in your coverage or your portfo-
lio. Accordingly, pick a common year to begin this exercise. It is highly useful to
have at least five years of historical data.
       Do we link the data directly from the income statement presentation work-
sheet? No. We put together our annual stacks on the ratios and valuations work-
sheet precisely so we could link to individual year values and drag and drop
values for multiple years. So let’s link to the annual EPS numbers in our annual
income statements on the R&V worksheet. Always driving the model-building
process is the search for balance between precision and ease of use.
                             Simple Average and Market-Weighted Comparisons    •   307



      We always open a single individual company workbook to start this exer-
cise; at most, you want to have just a few workbooks open during the linking
process. Under the earnings heading and in the first-year historical cell for the
specific company, link to the appropriate EPS measure (GAAP or pro forma) for
that year. Figure 15.2 shows our historical and forecast EPS for the peer group,
along with five-year (2004–2008) compound annual growth rate (CAGR). We’ve
also included EPS growth in our illustration; at the bottom of each EPS growth
column you can see both simple average and weighted average growth.
      A word about process. When you link, you’ll notice that the linked value
shows a dollar sign after the host cell column and another dollar sign after the
host cell row. Remove these dollar signs; otherwise, when you drag and drop
these values, you’ll be replicating the same cell contents. Once the dollar signs are
removed, drag and drop the annual EPS links up to the most recent full historical
year; in our example, that year is 2008.
      Let’s assume we’ve now dragged and dropped the EPS values for all compa-
nies in our composite up to the latest historical year, which is 2008 in our case.
In the next column, let’s calculate five-year CAGR for each stock as well as for the
composite. Many CAGR formulas will work in Excel. In the example, the simple
CAGR formula we use— (L15/H15)^(1/5) 1—shows Cisco’s pro forma earn-
ings grew at a compound annual growth rate of 15.5% for the five-year span
between fiscal 2004 and fiscal 2008.
      The trouble with a CAGR calculation is that it cannot handle a negative
number in any of the inputs. Accordingly, for a few companies in our earnings
grid, we’ve had to truncate the period to dodge opening-period losses. This is an
ongoing problem; when we switch the five-year period to 2005–2009, we’ll have
to deal with lots of losses from the difficult 2009 year. So while we’ve titled this
column “5-Year CAGR,” it is not purely that. Remember, though, we are seeking
a general trend. We can see that the general trend in EPS for the group was an
average CAGR in the very low double digits over the 2004–2008 span. That’s not
a bad number; but, given the imperfections, we treat this only as a broad trend
indicator and not as a definitive number.
      Now we need our two-year EPS projections. Assuming 2008 historical was
in column L and the CAGR in column M, link to 2009 estimated EPS in column
N. Remember, if you copy and past this link across two columns (including the
CAGR column), the source file will give you a value two columns over. So if you
copy and paste or drag and drop, adjust to ensure that the 2009 EPS value is in
the 2009 column. Drag and drop to get the 2010 value.
      The next set of earnings-related values is earnings growth. But didn’t we just
do that in the CAGR column? There is a distinction; we derive annual EPS growth
Figure 15.2

Earnings in the communications equipment universe have been “feast or famine” over the years, with strong midquarter results book-ended by the technology-
centered catastrophe in 2001–2002 and the much broader economic recession in 2007–2008. In our truncated view, the worst of that earlier period is mercifully
hidden.

Comparison Worksheet
Communications Equipment
                                                                                      Earnings
EPS                         1998        1999               2004        2005         2006       2007         2008      5-YearCAGR       2009E        2010E
ADC Telecom                     2.02        0.80               0.15        0.88         0.89        1.18         1.12        49.6%          0.27        0.49
Adtran                          0.51        0.66               0.93        1.29          1.14       1.17         1.32          7.3%         1.29        1.47
Alcatel                         3.10         1.19              0.59        0.80         0.28       0.40        (0.11)         -9.2%       (0.07)        0.21
Big Band                                                                                 0.16      0.01         0.34         28.7%
Ciena                           0.22        0.02              (2.15)       (1.11)       0.32       1.41          1.14        53.0%        (0.33)        0.32
Cisco Systems                   0.29        0.36                0.76        0.92         1.12      1.34         1.56          15.5%         1.35        1.41
Ericsson                                                        0.82        1.02         1.12      1.06         0.59          -6.3%         0.65        0.82
Extreme Networks                           (0.02)             (0.02)         0.13        0.14      0.02          0.14          2.1%         0.10        0.19
Corning Inc.                     0.54        0.67               0.46        0.84         1.12      1.40         1.54          27.5%          1.17       1.31
JDS Uniphase                     0.92        1.47             (0.29)      (0.50)       (0.10)      0.29          0.51         12.1%         0.23        0.27
Juniper Networks               (0.24)      (0.03)               0.44        0.72        0.73       0.87          1.18        63.6%          0.85        1.09
Motorola                         0.23        0.62               0.91         1.13        1.25      0.24         0.02        -52.2%          0.01        0.25
Nokia                            0.36        0.56               0.81        1.06         1.33      2.14         1.99          19.6%         0.89        1.42
Polycomm                         0.23        0.42               0.70        0.80        1.09       1.37         1.50          16.4%         1.26        1.66
Sycamore Networks                                             (0.15)      (0.04)         0.11      0.09         0.05        -24.5%        (0.08)        0.03
Tellabs                         0.98         1.32               0.61        0.56        0.57       0.24         0.24         -17.4%         0.29        0.35
                                                                                                                             11.6%
      Figure 15.2 (continued)


      Comparison Worksheet
      Communications Equipment
                                                                              Earnings
      EPS Growth                 1998       1999       2004        2005      2006        2007       2008        5-Yr Avg     2009E        2010E
      ADC Telecom                 -560.2%     -60.4%    -144.5%     496.0%        1.9%     32.2%       -4.8%         76.2%     -75.7%
      Adtran                      -460.2%      27.5%       22.3%     39.3%      -11.7%      2.8%       12.5%         13.0%      -2.3%       14.1%
      Alcatel                                            -147.8%     36.5%     -65.2%      43.1%     -126.4%        -51.9%     -29.3%
      Big Band
      Ciena                       -60.2%     -92.1%      -20.0%     -48.4%   -128.6%      342.9%      -19.2%         25.3%    -129.1%
      Cisco Systems                           22.7%       27.6%      21.2%      21.6%      20.2%       16.2%         21.3%       -13.1%       4.2%
      Ericsson                                                       24.7%       9.8%       -5.1%    -44.4%          -3.7%         9.7%     25.8%
      Extreme Networks                                   -88.5%                 10.6%     -85.6%         0.0%       -40.9%         0.0%     85.4%
      Corning Inc.                 -9.9%      24.4%      357.2%      83.4%      33.5%      25.2%       10.0%        101.9%      -24.3%      11.8%
      JDS Uniphase                 43.7%      60.0%       -77.6%     68.3%     -79.8%                   77.3%        -3.0%      -55.6%       17.8%
      Juniper Networks                       -85.5%      203.1%      63.7%       1.3%      20.0%       34.4%         64.5%       -27.5%     27.6%
      Motorola                    -70.0%     173.6%      133.6%      24.7%       9.9%     -80.8%     -90.6%          -0.7%      -54.2%
      Nokia                        30.0%      57.5%        -6.7%     29.7%      26.0%      60.8%        -6.9%        20.6%      -55.3%      59.6%
      Polycomm                    128.9%      82.8%        74.3%     14.7%      35.9%      25.1%         9.7%        31.9%      -15.7%      31.4%
      Sycamore Networks                                  -25.6%     -73.9%   -392.8%      -16.4%     -48.5%        -111.5%    -262.5%
      Tellabs                                35.2%      -415.8%      -9.3%       2.1%     -58.2%        -0.6%       -96.3%        24.5%     20.5%
      Simple Average             -119.8%     22.3%         -7.7%    55.0%     -35.0%      23.3%      -12.1%           3.1%     -47.4%       29.8%
      Weighted Average                                   52.2%      32.3%      15.3%      20.3%        -4.3%                   -23.8%       17.9%
309
310   •   Relational Valuation: The Industry Matrix Workbook and Peer Derived Value



in this grid, not five-year growth. You can use a simple percentage change for-
mula such as, (F15/E15) 1 to determine annual change for Cisco. Drag and
drop both across and up and down to fill out the grid. Note that the five-year
average of individual growth rates results in a value (3%) different from the five-
year CAGR (11%) calculated above. The CAGR calculation, remember, only
touched the five-year beginning and end points.


Revenues
We’re going to set up a grid of individual company revenues that will be very
similar to the earnings grid; that is, we will link to individual revenues for the
companies, and we will calculate annual growth rates. Mercifully, the revenue
tallies are all positive numbers, meaning we won’t have to leapfrog bad data
points. Look out for immature companies that may not have a five-year public
reporting history. For these companies, we use the appropriate truncated period,
or we can exclude them. Again, we seek the general trend.
      Simple average revenue growth for this group averaged 11.6% annually for
the preceding five years (2004–2008). That is a very healthy number. It is indica-
tive not just of cyclical recovery but of the secular trends at play among global
citizenry: more digital media, more mobile broadband, and an always-connected
lifestyle.
      When we compile the individual revenue growth rates for each year for each
company and use them to calculate five-year average growth, we get a 16% annual
growth rate for the five-year period. That does not square with the 11.6% cited
above. Coming out of the 2001–2002 recession, industry growth was strong for
several years. By 2008, of course, the world had changed; anyone doing business
late in 2008 had to wonder if the world would ever be what it once was. That bad
year only had a one-fifth influence on the individual average growth rates. The
CAGR calculation, by contrast, only touched the 2004 beginning point as well as
the 2008 end point; so its view is distorted. This is a reminder about the limita-
tions of unadjusted CAGRs and the main reason we use ordinary least squares
(OLS) regression to smooth growth rates in most places.
      What have we learned from our earnings and revenue composites? From
earnings, the value of the historical analysis has been limited by the many adjust-
ments we needed to make to avoid those loss years that skew CAGR calculations.
Looking at 2009 in the midst of its unfolding, we learned that communications
equipment earnings were having an awful time, down 49% on a simple average
basis and down 23% on a cap-weighted basis. We are anticipating double-digit
recovery in 2010, however, with the simple average composite somewhat better
                            Simple Average and Market-Weighted Comparisons   •   311



than the weighted composite. That tells us that smaller aggressive growth com-
panies are situated to grow faster than their larger rivals—a typical midcycle
phenomenon.
      Revenues, as noted, are always easier to work with. The five-year revenue
growth of 11% on a CAGR basis and 16% on an average-of-averages basis
were impressive—three to four times the average U.S. GDP growth in that
span. Looking forward, the sum of all our individual analysis suggest that
revenues will decline 16% in 2009 on a simple basis and 12% on a weighted
basis. Things could have been worse; a glance back across the grid shows that
the simple average revenue decline as 34% in 2002. During that earlier finan-
cial crisis, technology was at the center of the catastrophe; in the 2008–2009
global recession, technology was merely a spear carrier while the financial
sector sang Götterdämmerung.
      The revenue rebound estimated for 2010 is fairly tame: a forecast 8% on a
simple basis and 6% on a weighted basis. But we see that in 2003 segment reve-
nues were still negative. When real recovery finally came in 2004, it was in the
double digits.
      The point of these observations is certainly not in the observations them-
selves, which are common knowledge. The point is that compilation of data
enables the analyst to not just quantify arithmetic trends but to discern broader
industry trends within the context of past trends.


Balance Sheet Metrics
Amid all our other priorities, we have likely scanted the significance of changes
in the accounts constituting working capital, and the cash flow implications of
such changes. CFOs certainly do not scant on this material. Read the transcript
of any earnings conference call and the CFO will devote time to the discussion
of days sales outstanding and the cash cycle. In the post-results Q&A, analysts
will pick at this data. And no wonder. Days spent in inventory may seem arcane,
but it reflects money tied up; the massive market caps of major corporations can
mask the scarcity sometimes of liquidity. Accounts receivable (A/R), accounts
payable (A/P), and inventories resonate in the gross margin and really across the
operating structure.
      While we could import significantly more operating and liquidity data
from our ratios and valuations page, we tend to focus on three areas: working
capital, days sales outstanding, and cash cycle. As noted, professional investors
are tightly focused on these metrics because they can tell us much about a com-
pany’s efficiency and execution.
312   •   Relational Valuation: The Industry Matrix Workbook and Peer Derived Value



       When I was first entering the analysis business, overstuffed current ratios
were in vogue. I took that as a given, not realizing that I’d started my analysis
career at the top of the market. As I experienced my first down cycle, I was mysti-
fied that the same senior analysts who had previously lauded companies for large
current ratios now complained about them. Their complaints were valid, in that
a fat current ratio in a down market likely means cash tied up needlessly in inven-
tories that might need to be written down; a sloppy collections experience; and
perhaps too many receivables at potentially insolvent companies about to be
blindsided by the recessionary freight train.
       In recent years, companies have come to value cash on the balance sheet
much more so than in the past. On that basis, unadjusted current ratios may
seem suspiciously high. As a reminder, we use trade working capital—which
kicks out cash and other things from the calculation—in our DFCF work. Trade
working capital is likely a better representation of the daily cash uses and needs
of the enterprise. In that regard, it may be the better metric for the industry
matrix.
       Companies continually strive to shorten their days sales outstanding (DSO)
and their cash cycle. Geographic mix is a bit of a wild card; sales terms are more
liberal (longer) in Europe and Asia. Deterioration in DSO and cash cycle can
actually be a sign of global growth. More significantly, no matter how effectively
companies streamline their procurement and global supply chains, they cannot
control the economic cycle and the end market. Inevitably, they will have built
apparatuses that are somewhat oversized and underutilized for demand in bad
times and somewhat undersized and strained in good times.
       Therefore, we analyze the trends in DSO and cash cycle with several things
in mind. Certainly, we look for progress over the year, both in absolute terms
(specific to historical trends) and in relation to the peer group. But we also seek
to analyze the varying performances in times of economic distress to see who has
a flexible supply chain and who doesn’t. One company that has had more than
its share of woes since inception, Alcatel-Lucent, appears to be doing well, on
these metrics at least. More than any of the company’s press releases, which have
been proclaiming the success of this integration for years, improvement in these
important balance sheet metrics can be read as a sign that integration in all the
out-of-view and nonoperating niches is finally being achieved.


Margins
The inputs on the company comparison worksheet are designed to help us deter-
mine an industry standard and to assess to what degree the individual companies
                             Simple Average and Market-Weighted Comparisons   •   313



vary from that standard. Not every input is equal; the cash cycle and DSO data
are of interest to a slim wedge of the investing public. Everyone, however, looks
at margins; they are simple, explicit, and easily understood. During company
presentations, CFOs spend lots of time heralding margin progress or explaining
margin shortfalls.
      On the company comparison worksheet, both the individual year averages
and the five-year average provide a clear industry standard. Of equal interest is
the individual company trend in margins. Of course, you can track the progress
in, say, gross margins, on the individual equity workbook. But it is always useful
to see the progress in context.
      It therefore sometimes makes sense to calculate the five-year CAGR in margins.
If a company has a 2008 gross margin that lags the group, some of the stigma is
reduced if this same company has expanded its gross margin at a rate exceeding that
of the peer group; the CAGR measure of margins can provide that information.
      The margins most typically measured in an industry matrix worksheet are
gross margins, operating margins, pretax margins, net margins, pro forma net
margins, and EBITDA (earnings before interest, taxes, depreciation, and amor-
tization) margins. As with earnings and revenues, we input as many historical
periods as we can for all equities, with a minimum of five years to enable an
historical basis. We calculate simple average and weighted average for each year,
along with simple average for the five-year period.
      Figure 15.3 shows the company grid organization, and it also provides some
interesting color on margins. Gross margins across this industry niche held up
better in 2008 than in 2002; despite the more proactive cost-cutting this time
around, pretax margins still sagged deeply from mid-decade levels.


Return Measures
The most common return measures captured on an industry-average basis on the
company comparisons worksheet are return on equity, return on assets, return
on capital, and return on invested capital. These are GAAP measures. Because of
the long-standing habit of technology companies to focus on non-GAAP earn-
ings, and investors’ complicity in this practice, we believe technology companies
place a low priority on their GAAP results.
      The net effect is that when ROE or ROA are analyzed for this group, they
tend to be surprisingly puny. That is particularly true of ROA. And that is likely
because boards and managements can’t squander the book value of assets the way
they decimate the book value of equity, through pricey and misguided acquisi-
tions and big-bath write-offs.
Figure 15.3

In the bad days of 2002, simple average gross margins for the communications equipment companies slipped under 20%. By 2008, companies had fully outsourced,
substituting variable cost for fixed costs; that helped hold up gross margins. Net margins did not fare so well, however.

Comparison Worksheet
Communications Equipment
                                                                                     Margins
Gross Margin            1998      1999     2000      2001     2002       2003     2004    2005       2006     2007      2008 5-Yr Avg 2009E 2010E
ADC Telecom               48%       47%      49%       30%      24%        38%      38%      35%       32%      33%       34%     35%    33%   34%
Adtran                    55%       51%      50%       45%      51%        56%      57%      59%       59%      59%       60%     59%    60%   60%
Alcatel                   27%       29%      29%       25%      29%        34%      38%      35%       37%      32%       34%     35%    34%   36%
Big Band                                                                                                        49%       61%     55%    58%   62%
Ciena                      50%       38%      44%       44%      -53%     24%       27%       32%      46%      46%       50%    40%     45%   47%
Cisco Systems              66%       65%      64%       60%       61%     70%       69%       67%      66%      64%       64%    66%     64%   65%
Ericsson                                                          29%     33%       46%       46%      41%      39%       36%     42%    37%   38%
Extreme Networks           37%      51%       52%      43%        47%     41%       51%       53%      54%      54%       57%     54%    57%   57%
Corning Inc.               40%      39%       43%      30%        14%     28%       37%       43%      44%      47%       47%    44%     37%   38%
JDS Uniphase                        52%       51%      34%         -5%    11%       23%       18%      28%      34%       39%     28%    39%   41%
Juniper Networks          -32%      56%       65%      62%        58%     63%       69%       68%      67%      67%       67%    68%     66%   69%
Motorola                   39%      40%       38%      32%        33%     33%       34%       32%      30%      27%       28%     30%    30%   31%
Nokia                      38%      38%       37%      37%        39%     41%       38%       35%      33%      34%       34%     35%    32%   33%
Polycomm                   50%      56%       59%      60%        59%     61%       63%       62%      62%      58%       58%     61%    57%   58%
Sycamore Networks                   25%       47%      16%      -134%      8%       35%       49%      50%      43%       46%     45%    40%   51%
Tellabs                    59%      60%       54%      46%        37%     36%       53%       45%      46%      35%       38%    44%     42%   41%
Simple Average             40%      46%       49%      40%        19%     39%       45%       45%      46%      45%       47%    46%    46%    47%
Weighted Average                                                          53%       54%       54%      52%      51%       51%           50%    51%
      Figure 15.3 (continued)


      Comparison Worksheet
      Communications Equipment
                                                                                            Margins
      Pretax Margin             1998     1999     2000     2001      2002      2003      2004    2005      2006     2007      2008 5-Yr Avg 2009E      2010E
      ADC Telecom                17.3%     6.3%    44.4%   -79.9%    -84.2%     -10.6%    -0.3%     7.2%     4.4%    10.8%      -2.6%    4% -40.7%        2.4%
      Adtran                     21.2%    21.0%    39.5%     6.1%      9.3%      22.4%    24.3%   29.9%     25.1%    24.2%      23.6%   25%  22.0%      22.8%
      Alcatel                     0.0%     0.0%     0.0%     0.0%      0.0%       0.0%     0.0%     1.7%     0.0%      2.3%       2.3%   1%     1.6%      1.5%
      Big Band                                                                                                      -13.7%        6.6%  -4%   -8.0%       8.4%
      Ciena                       18%       1%      14%    -106%      -246%     -137%     -265%                      10.6%        4.3%   7% -88.0%       -4.0%
      Cisco Systems               25%      24%      16%      14%         12%       27%      32%   32.4%    33.4%     32.0%      30.5%   32%  26.2%      28.6%
      Ericsson                                                          -16%      -10%      22%    22%      20%         16%       8.2%  18%     8.8%     11.4%
      Extreme Networks           -59%        0%     12%    -19%        -52%       -17%       0%    4.3%     3.6%     -3.5%        2.9%   2%     1.6%      4.3%
      Corning Inc.                12%       14%     10%    -97%        -97%      -20%        6%   16.6%    19.8%     25.5%      21.4%   18%     9.7%     10.5%
      JDS Uniphase                28%       32%    -37% -1568%        -746%     -123%      -20%                       -1.7%      -1.4%  -8% -62.5%       -2.4%
      Juniper Networks                      -6%     34%     13%          -4%       11%      19%   24.3%    -38.0%    18.0%      20.8%    9%   13.6%      19.2%
      Motorola                     2%        6%      6%    -19%         -12%        6%      10%    17.7%    10.8%     -1.1%     -8.7%    6%    -1.0%      2.5%
      Nokia                       18%       19%     19%     11%          16%       18%      16%   14.5%     13.9%    16.2%        9.8%  14%     5.3%     10.5%
      Polycomm                    15%       21%     18%      1%          10%        8%      11%   16.5%     15.1%      8.8%       9.1%  12%     9.3%    13.2%
      Sycamore Networks                   -172%     15%    -71%       -583%     -150%     -103%                       -7.9%       0.4%  -4% -45.5%        5.9%
      Tellabs                     34%       35%     32%     -5%           1%        4%      19%    11.3%    13.9%      3.7%    -55.1%   -1%     6.8%       7.7%
      Simple Average            11.0%      0.1%   15.9% -137.1%     -119.5%    -24.9%    -15.2%   16.5%    10.2%      8.8%       4.5%  8.2%  -8.8%        8.9%
      Weighted Average                                                            12%       20%     24%    20.9%    22.7%      18.8%         14.7%      18.5%

                                                                                                                                                   (continued)
315
Figure 15.3 (continued)


Comparison Worksheet
Communications Equipment
                                                                                         Margins
Net Margin                1998     1999       2000      2001      2002      2003      2004    2005        2006     2007       2008 5-Yr Avg 2009E     2010E
ADC Telecom                11.2%     3.6%      26.4%    -53.6%   -109.9%     -9.9%      3.3%     5.4%       7.0%     10.4%      -2.9%    5% -40.7%       1.8%
Adtran                     14.1%    13.9%      26.1%      4.5%      7.2%     15.5%     16.6%    19.7%      16.6%     16.0%      15.7%   17%   14.8%     15.1%
Alcatel                    -0.8%    -0.9%      -0.4%      0.1%      8.0%      3.6%      0.5%     0.1%      -0.1%     -0.6%    -30.9%    -6%   -3.4%      1.9%
Big Band                                                                                                            -14.4%        5.3%  -5%   -6.0%      6.8%
Ciena                      10.1%     1.9%      9.5%      -111.9% -276.7%    -136.5%   -265.1%                        10.6%        4.3%   7% -88.0%      -4.0%
Cisco Systems              13.5%    13.9%      6.7%        -4.5%     5.6%     19.0%     22.5%    23.1%    21.8%      21.7%     20.4%    22%   17.0%     17.4%
Ericsson                                                           -13.0%     -9.2%     14.4%    16.0%    14.9%      11.8%        5.6%  13%    6.2%      8.0%
Extreme Networks          -59.1%      -1.6%     7.7%      -14.0%   -41.7%    -54.3%      -0.2%    3.4%     3.1%       -4.1%       2.3%   1%    0.8%      3.2%
Corning Inc.               11.0%     10.7%      5.8%      -87.5% -65.9%       -2.9%      -9.1%   13.2%    35.9%      36.9%     46.0%    25%   31.5%    34.6%
JDS Uniphase               18.3%     21.2%    -41.2%   -1563.9% -745.7%     -125.3%     -17.5%                        -1.9%      -1.6%  -7% -62.1%      -1.9%
Juniper Networks                     -8.8%     22.0%       10.0%    -4.5%      7.7%     12.5%     17.1%   -26.9%     12.7%      14.6%    6%    5.8%    13.8%
Motorola                    1.6%       4.3%     4.5%      -13.4%    -8.6%      4.0%       7.0%   12.5%      7.6%     -0.3%     -14.1%    3%   -0.7%      1.6%
Nokia                      12.6%     13.0%     13.0%        7.1%    11.3%     12.2%     10.8%    10.6%     10.5%     14.1%        7.9%  11%    4.2%       7.1%
Polycomm                   13.3%     14.4%     11.4%       -5.0%     7.3%      5.5%       7.9%   10.7%     10.5%       5.9%       7.1%   8%    6.9%      9.6%
Sycamore Networks                  -172.0%     10.3%      -74.7% -582.5%    -149.7%   -103.2%                        -8.5%       -0.1%  -4% -45.3%       5.8%
Tellabs                   23.2%      23.7%     22.0%       -4.0%     1.7%      4.5%      17.9%    9.3%     9.5%        3.4%   -53.8%    -3%    4.1%      5.1%
Simple Average               6%         -4%       9%      -136% -120%         -28%       -19%     12%        9%        7.1%      1.6%    6%  -9.7%       7.9%
Weighted Average                                                                 9%       13%     17%      16%       17.7%     14.1%         11.3%     14.2%
                              Simple Average and Market-Weighted Comparisons     •   317



      For our subset of the technology universe, return on assets for the five-year
historical period is a slight 4.3%. Negative outliers distort the simple average
ROA for 2009, rendering it negative; this sort of number underscores our reliance
on cap-weighted metrics. By 2010, companies presumably will be benefiting from
a roaring recovery; and we, like most analysts, have baked nothing but sunshine
into our year-out estimates, even with the knowledge that things will go awry.
Even so, the simple average ROA forecast for 2010 is a paltry 5%—more akin to
an insurer’s ROA.
      The absolute numbers in ROE are somewhat better, but it has little to do
with the numerator and much to do with the denominator. We note a significant
variance between simple average ROE and weighted ROE; for most of the histori-
cal period, nearly all the returns on equity were coming from two companies,
Nokia and Cisco. Now it’s a bit more democratic, and in fact Nokia’s contribution
is hurting. Off the five-year simple average of 7%, ROEs are on track to slip to a
negative 16% in 2009 and then rebound to a positive 9% in 2010. But the main
driver of this gain is all the impairment losses that have decimated stockholders’
equity, making it that much easier to earn return on it.


Valuations
We’ve saved, if not the best, than certainly the area of most intense investor interest
for last. The historical comparables we calculated on the individual equity work-
books helped to determine dollar value of the individual asset; they have more to
do in this category. But first let’s simply gather the averages for these key metrics—
price to earnings, sales, book value, cash flow, and relative P/E—to see what they
say about the industry averages and the various individual equity variations.
      Figure 15.4 illustrates a positive fact about which many investors are evi-
dently unaware: even as the quality of revenues, earnings, and balance sheets
among technology companies have improved, price/earnings ratios have contin-
ued to moderate.
      The five-year trend in P/Es in contrast to the forward P/Es demonstrates a
watershed change in perception of the technology shares. Let’s walk through the
P/E history of this particular group for a moment to demonstrate the value of this
kind of data gathering. The P/Es for the group were hysterical back in 2000 and
2001, as investors spoke of “new paradigms” and some enlightened investors
declared technology profits a sign of timidity. Many stocks followed the new
paradigm from the $100-plus level to the penny stock sheets, and the reputation
of the tech companies—reckless, detached from reality, and an abscess in the
portfolio—was set.
Figure 15.4

Valuations for technology companies in general have been coming in for years. The formerly wild and crazy group is now solvent, cash rich, low on debt—and the
market keeps driving down sector P/Es.

Comparison Worksheet
                                                                                        Valuations
Price to Earnings     1998      1999      2000      2001       2002       2003       2004     2005        2006       2007       2008       5-Yr Avg 2009E 2010E
ADC Telecom             176.2     626.3     170.4     64.8      (15.2)     (50.0)      121.1     22.3       22.3        14.7        10.5          38.2  31.8  17.7
Adtran                   24.7      23.8      18.2     54.6        27.5       33.4       29.1     18.9        21.9      20.9         14.5          21.1  17.6 15.4
Alcatel                     -         -         -        -       (2.7)       (7.5)      24.6     16.5        47.7      28.2       (31.1)          17.2     - 15.9
BigBand                                                                                                                             15.1          15.1 52.8   11.4
Ciena                   111.0    805.5      253.8      15.0      (17.6)    (14.3)     (15.7)    (14.4)        91.4      23.9       22.2           21.5     -  37.9
Cisco Systems            44.0     62.4      101.3     100.5        42.8      28.9       30.8      20.5        17.5       17.5       17.9          20.8 14.2  14.9
Ericsson                                                        (14.9)     (14.1)       16.0      15.8       15.6       13.3        16.4          15.4 14.3   11.4
Extreme Networks                (894.1)     183.9      531.0     260.2     (41.2)                 44.8       34.4                  26.3           35.2 22.0  13.4
Corning Inc.             21.2      43.4      59.9      150.0      (9.9)      74.9       24.9      18.8        21.0      16.3        11.4          18.5 13.5   12.1
JDS Uniphase             49.7      68.1     196.8    1,182.9    (38.9)      (17.9)   (116.7)    (40.2)    (226.7)      59.6        25.4         (59.7) 26.0  22.1
Juniper Networks              (1,083.2)     259.0       91.9          -      90.9       55.6      32.7       23.2       31.7        19.7          32.6 28.7  22.5
Motorola                 75.5      51.5      48.1     (58.6)       88.3      25.2        17.2      17.4      18.0      75.1       428.0          111.1     - 28.1
Nokia                       -         -         -       36.3       20.5       18.4      19.6      15.4       15.5       14.1        12.8          15.5 14.5    9.1
Polycomm                 28.8      41.6      81.2       37.8       30.8      35.7       29.9      23.0       22.8       21.8        13.8          22.3 18.0  13.7
Sycamore Networks                         1,679.3   (469.6)       (9.7)    (15.8)     (25.4)    (95.6)       38.7      34.6        76.2            5.7     -  91.3
Tellabs                  26.2      41.6      33.1      116.4          -          -      15.0       16.1      22.8      39.8         21.9          23.1 22.4  18.6
Simple Average           47.2   (100.5)     131.6      231.9      36.1       18.6        6.9      13.4         3.3     30.4        59.4          23.3  15.1  18.5
Weighted Average                                                             27.3       26.1      18.6       18.8      21.1        40.0          24.4  13.8  14.8
      Figure 15.4 (continued)


      Comparison Worksheet
                                                                                            Valuations
      Relative P/E         1998        1999      2000      2001      2002      2003      2004     2005       2006      2007       2008       5-Yr Avg    2009E      2010E
      ADC Telecom             7.36       24.39     6.75       2.48    (0.74)         -      7.00     1.40       1.53      0.87       0.51          2.26     1.60        1.10
      Adtran                  1.03        0.93     0.72       2.09      1.35      1.92      1.71     1.21       1.52      1.31       0.67          1.28     0.88       0.98
      Alcatel                 0.42        1.07      1.46    (0.57)    (0.13)    (0.43)      1.48     1.06       3.31      1.71     (1.77)           1.16        -       1.12
      BigBand                                                                                                                        0.86          0.86     2.65       0.76
      Ciena                     4.64     31.37     10.06     0.58     (0.82)    (0.83)    (0.91)   (0.91)       6.28      1.45       1.32          1.45    (2.19)      2.53
      Cisco Systems             1.96      2.51      3.96     3.94       1.85      1.62      1.73     1.22       1.15       1.10      0.84          1.21     0.78       0.88
      Ericsson                                                        (0.73)    (0.81)      0.94     1.01       1.08      0.79       0.75          0.92     0.72       0.75
      Extreme Networks                 (34.83)      7.29    20.36      12.50    (2.45)               2.63       2.04     11.93       1.25          4.46     1.06       0.78
      Corning Inc.              0.89      1.69     2.37       5.75    (0.49)      4.30      1.46     1.20       1.46      0.96       0.65           1.15    0.68       0.80
      JDS Uniphase              2.08      2.65      7.80    45.36     (1.66)    (1.01)                                    3.68        1.31         2.50      1.46      1.28
      Juniper Networks                            10.26       3.52         -      5.22      3.27     2.09       1.54      1.81       0.98          1.94      1.44       1.47
      Motorola                  3.09      1.96     1.99     (2.11)      4.41      1.45      1.01      1.11      1.25      4.45      19.65          5.50         -      1.86
      Nokia                     1.05      1.47     2.21       1.39      1.00      1.06      1.15     0.99       1.07      0.95       0.64          0.96     0.73       0.59
      Polycomm                  1.20      1.62     3.22       1.45      1.51      2.05      1.76     1.47       1.58      1.25       0.63          1.34     0.90       0.89
      Sycamore Networks                           66.52    (18.01)    (0.48)    (0.88)    (1.56)                          2.47       4.84          1.92         -      8.70
      Tellabs                   1.09     1.62       1.31      4.46         -         -      0.88     1.04      1.59       2.21       1.25          1.39      1.13       1.21
      Simple Average            2.25     3.04      8.99      5.05       1.17     0.75      1.53      1.20      1.95      2.46        2.15          1.89     1.88       1.61
      Weighted Average                                                            1.56     1.53      1.17      1.27       1.31       1.79          1.29     1.22       0.94
319
320   •   Relational Valuation: The Industry Matrix Workbook and Peer Derived Value



      But the industry did not stand still, and elements of the sector’s immaturity—
garage start-ups, it turned out, had no defined-benefit pension obligations—were
eventually recognized as strengths. Low debt, lots of cash, balanced global customer
bases: quietly, technology turned respectable. Meanwhile, the stocks with lampshades
on their heads at the party turned out to be—the money center banks.
      The five-year average P/E for this formerly wild group came down to the
low 20s, and former 2.0 relative P/Es were suddenly in the 1.25 range. But it is the
weighted P/Es for 2009 and 2010 that reveal how investors really feel about these
stocks. Technology is still massively transformative, the sudden revelation of
which led to their hyperinflated status at the turn of the century. But massively
transformative is woven into the investment story now, and investors tend to treat
the familiar with . . . familiarity.
      Once again, the point of this discussion is not so much to explain the nature
of investors’ changing perceptions of a specific market niche, but to illustrate the
ways in which collected data of this nature substantiates and quantifies such
observations. We can also trace the trajectory of investors’ perceptions of indi-
vidual companies. Juniper, arguably the most successful second banana in the
technology industry, routinely enjoyed P/Es in the 90s; even as the industry
steadied in the 2003 and 2004, its P/Es remained well above industry trend. Grad-
ually, its P/Es have come down closer to the industry average, while still retaining
a premium. Later we’ll talk about quantifying this observation, but for now it is
sufficient to track the change.


Segment Worksheet
We won’t spend too much time on this worksheet. It is principally a compilation
of the “greatest hits” from the company comparisons worksheet and, thus, makes
a handy tear sheet for client handouts, presentations, “squawk-box” calls, or more
formal client visits. A great deal of investment decision making comes down to
the data on this worksheet, which includes EPS and EPS growth, revenue and
revenue growth, P/Es, relative P/Es, and enterprise value/EBITDA.
      The distinguishing feature of this worksheet compared to the company
comparison sheet is that it breaks out the companies principally by their major
end markets. The communications equipment world, like most technology
niches, is now full of blurred boundaries; the days of distinct customer silos are
over. Companies that principally serve the service providers (e.g., Verizon and
AT&T) also serve enterprise or business customers, and former enterprise IT-
only companies such as Cisco now have multibillion-dollar carrier businesses.
                              Simple Average and Market-Weighted Comparisons     •   321



       With information processing and networking now colliding in the data cen-
ter as we write, the final technology silos may be coming down. Still, grouping
and analyzing the principal valuation and growth data by subsectors is useful for
understanding what the market is really thinking.
       Figure 15.5 provides a stark view, in this particular industry, of how difficult
it is to serve the service provider market. All the enterprise-centric companies are
currently profitable, as are the component companies. Among the vendors primar-
ily serving the carrier market, several are unprofitable—chronically so.


Quarterly Trends Worksheet
The matrix workbook is the logical place to aggregate key quarterly data. There
are four quarters in the year, and if your industry matrix has been around for a
while as this one has, it can become overloaded with quarterly pages more quickly
than you’d think. (But then I’ve reached that age, to paraphrase the novelist
Richard Powers, where my birthday seems to come around every six weeks or so.)
Make a point of offloading the quarterly sheets before they build up too much
and Excel gets cranky.
      The market-weighting function really displays its worth in times of transi-
tion, when investors are scouring the skies for breaks in the clouds or more black
cumulus. Figure 15.6 shows the revenue trends for 2Q09, a period for which
investors knew they were getting negative year/year comps—and in which the
investor focus was most intense on a sequential basis. The sequential revenue gain
shown in cell G21—4% quarter-over-quarter growth on a simple average basis—
tells some of the story. But in our view, that 4% is mainly reflective of what
companies pushed out into the market.
      The cap-weighted sequential change in cell I22, by contrast, tells less what
the company did and much more what their customers did. Those customers
grew demand by 8%, not the misleading 4% of the simple average. The industry
really grew 8% sequentially, not 4%. This “better” information is one reason we
dedicate an entire worksheet (or two) to market-weighting all our data.


Weightings Worksheet
Market Cap Weighting of Data Points: An Overview
One benefit of the industry matrix over public or purchased data is the ability to
manage and tailor data to your specifications based on the market capitalization
Figure 15.5

The sector comparison worksheet has a “greatest hits” flavor. The huge amount of individual company data cooked down on the comparisons worksheet is cooked
down further here, providing a snapshot for clients and a cheat sheet for numbers-numbed analysts.

Communications Equipment Universe
                                        2004   2005     2006     2007     2008 2009E 2010E 04 EPS 05 EPS 06 EPS 07 EPS 08 EPS 09E EPS 10E EPS
COMPANIES                Tckr   Price   EPS    EPS      EPS      EPS      EPS   EPS   EPS Growth Growth Growth Growth Growth Growth Growth
Carrier
Adtran                   ADTN   22.66    0.93    1.29    1.14      1.17     1.32    1.29   1.47   22.3%     39.3% -11.7%        2.8%  12.5%  -2.3%     14.1%
Alcatel-Lucent           ALU     3.37    0.59    0.80    0.28     0.40    (0.07)    0.21      - -147.8%     36.5% -65.2%      43.1% -126.4% -29.3%      0.0%
Ciena (b)                CIEN   12.29   (2.15) (1.11)    0.32     1.41       1.14 (0.33)   0.32 -20.0%     -48.4% -128.6%   342.9% -19.2% -129.1%       0.0%
Ericsson                 ERIC    9.30    0.82    1.02    1.12     1.06      0.59    0.65   0.82     0.0%    24.7%    9.8%      -5.1% -44.4%   9.7%    25.8%
Motorola                 MOT     7.03     0.91   1.13    1.25     0.24      0.02    0.01   0.25 133.6%      24.7%    9.9%   -80.8% -90.6% -54.2%        0.0%
Nokia                    NOK    12.95     0.81   1.06    1.33     2.14      0.89    1.42      -    -6.7%    29.7%   26.0%     60.8%   -6.9% -55.3%    59.6%
Sycamore Networks (c )   SCMR    3.08   (0.15) (0.04)    0.11     0.09      0.05 (0.08)    0.03 -25.6%     -73.9% -392.8%    -16.4% -48.5% -262.5%      0.0%
Tellabs                  TLAB    6.59     0.61   0.56    0.57     0.24      0.24    0.29   0.35 -415.8%     -9.3%    2.1%   -58.2%    -0.6%  24.5%    20.5%
Average                                                                                          -57.5%      2.9% -68.8%     36.1% -40.5% -62.3%      15.0%

Components
ADC Telecom (a)          ADCT     8.65   0.15   0.88     0.89     1.18     1.12    0.27    0.49 -144.5% 496.0%       1.9%   32.2%    -4.8%   -75.7%    0.0%
Corning Inc.             GLW     15.79   0.46   0.84      1.12    1.40     1.54     1.17   1.31 357.2%   83.4%      33.5%   25.2%   10.0%    -24.3%   11.8%
JDS Uniphase (d)         JDSU     5.85 (0.29) (0.50)    (0.10)    0.29     0.51    0.23    0.27 -77.6%   68.3%     -79.8%    0.0%    77.3%   -55.6%   17.8%
Average                                                                                          45.0% 215.9%     -14.8%    19.1%   27.5%    -51.9%   9.9%
      Figure 15.5 (continued)


      Communications Equipment Universe
                                               2004    2005    2006    2007    2008 2009E 2010E 04 EPS 05 EPS 06 EPS 07 EPS 08 EPS 09E EPS 10E EPS
      COMPANIES                 Tckr   Price   EPS     EPS     EPS     EPS     EPS   EPS   EPS Growth Growth Growth Growth Growth Growth Growth
      Enterprise
      Cisco Systems (c )       CSCO     21.04   0.76    0.92    1.12    1.34    1.56   1.35   1.41     27.6%   21.2%     21.6%    20.2%   16.2%    -13.1%     4.2%
      Extreme Networks (d)     EXTR      2.57 (0.02)    0.13    0.14    0.02    0.14   0.10   0.19   -88.5%              10.6%   -85.6%    0.0%     85.4%     0.0%
      Juniper Networks         JNPR     24.47   0.44    0.72    0.73    0.87    1.18   0.85   1.09   203.1%    63.7%      1.3%    20.0%   34.4%    -27.5%    27.6%
      Polycomm                 PLCM     22.71   0.70    0.80    1.09    1.37    1.50   1.26   1.66    74.3%     14.7%    35.9%    25.1%    9.7%    -15.7%    31.4%
      Average                                                                                         54.1%    33.2%     17.4%    -5.1%   15.1%      7.3%   15.8%

      Total
      Simple Average                                                                                  -7.7%    55.0%    -35.0%   23.3%    -12.1%   -47.4%   29.8%
      Weighted Average                                                                               52.2%     32.3%     15.3%   20.3%     -4.3%   -23.8%   17.9%

      (a) -- September fiscal years
      (b) -- October fiscal years
      (c ) -- July fiscal years
      (d) -- June fiscal years
323
Figure 15.6

In periods of crisis and transition, as in early 2009, investors want to know how much the overall market grew. Whereas cell G21 tells us how much the companies
on average pushed out into the market, cell I22 tells us how much the end market pulled in, and in our view that is the truer demand indicator.

2Q09 Revenue Trends
(*) - estimated                          Calendar 2Q09          Market                                 Calendar 2Q09    Market
Company                        TCKR      Y/Y Revenue %       Capitalization Times Market Cap           Q/Q Revenue % Capitalization Times Market Cap
ADC Telecom (*)                ADCT               -25.9%                 1,013           (262)                    2.9%          1,013                30
Adtran                         ADTN                 -7.4%               1,440            (106)                   10.1%          1,440              146
Alcatel                         ALA                 -4.8%               7,737            (370)                    8.5%          7,737              660
Big Band                       BBND                 -9.3%                 258             (24)                  -11.1%            258             (29)
Ciena (*)                      CIEN               -38.4%                1,284            (493)                    8.3%          1,284              107
Cisco Systems                  CSCO                -17.6%             129,422        (22,840)                     4.1%        129,422            5,360
Ericsson                        ERIC                 7.4%              29,695            2,209                    5.2%         29,695            1,541
Extreme Networks               EXTR                -17.3%                 296             (51)                    5.3%            296                16
Corning Inc.                    GLW                -21.6%              25,120         (5,433)                    41.1%         25,120           10,312
JDS Uniphase (*)               JDSU               -28.5%                1,334            (380)                   -0.5%          1,334               (7)
Juniper Networks               JNPR               -10.5%               13,615          (1,435)                    2.9%         13,615              395
Motorola                       MOT                -32.0%               15,934         (5,096)                     2.3%         15,934              374
Nokia                          NOK                -24.6%               48,926        (12,050)                     7.1%         48,926            3,488
Polycomm                       PLCM               -15.1%                1,981            (298)                    2.4%          1,981                47
Sycamore Networks (* )         SCMR                  4.1%                  874              36                  -31.7%            874            (277)
Tellabs                        TLAB               -10.8%                2,705            (292)                    6.6%          2,705              177
                                                                      281,635        (46,887)                                 281,635          22,340
Simple Average Change                              -15.8%                                                        4.0%
Cap-Weighted Change                                                                       -16.6%                                                 7.9%
                             Simple Average and Market-Weighted Comparisons     •   325



weighting of ratios, valuations, averages, and other data points. When assessing
the peer group, we regard market-weighted outputs as inherently more valuable
and “truer” than simple average outputs.
      If you were to ask a portfolio manager (PM) how he or she was doing, and
the PM replied by providing equal-weighted return, you’d perceive the reply to
be disingenuous. Who cares about a simple average of asset returns? Only the
size-weighted return captures the true performance and, along with the change
from opening assets under management (AUM), provides insight into how much
money the PM made or lost.
      As an analyst within a single industry or niche, you’ll likely find yourself
covering widely disparate companies based on their size and available market.
The smaller companies are apt to throw off lots of “outlier” data points relative
to larger and more staid enterprises; it’s tough for small companies to stay in
coverage if they don’t meaningfully outperform. But if we simple-average these
outlier performances, we likely overstate the strength of the available market in
the up cycle and understate it during the down cycle.
      In calendar 2Q09, Corning grew its revenues 41% sequentially; a much
smaller rival experienced 32% sequential contraction. Corning’s market cap is
26 times that of its smaller rival; should we allow the mirror-image perfor-
mances to cancel each other out? The reason we are interested in market cap–
weighted performance is that it gives a truer picture of activity in the available
market, not just what the individual companies are doing. During 2Q09, the
sequential revenue growth rate for 16 communications equipment companies
was 3.8% on a simple average basis; but the market-weighted sequential growth
was 8%. As these 16 companies represent a substantial portion of the commu-
nications equipment market, the 8% growth best represents the industry
trend.
      It’s true that Corning’s market cap is 26 times that of its smaller rival; were
2Q09 revenues 26 times higher? Actually, Corning’s sales were more like 15 times
higher than those of the small rival. So for this calculation, should we revenue
weight the growth instead of cap weighting it?
      Ideally, yes; but remember, we work in an office—maybe a cubicle, maybe
a corner office—not an ivory tower. While we won’t discourage the practice out-
right, using line-specific weights (e.g., revenues and EPS) to calculate line-specific
returns is cumbersome and will clutter an already crowded industry matrix
workbook. And, once again, market cap weighting is the industry convention,
and asset values implicitly encompass this verity.
326   •   Relational Valuation: The Industry Matrix Workbook and Peer Derived Value



Market Cap Weighting: The Process
You’ve heard us say this one before: there are several ways to market weight
returns, margins, ratios, and other data points. No matter which method you use,
it will be cumbersome and have lots of footprint (i.e., it takes up lots of space).
For each year and data series to be weighted, using our method you will need two
columns. On our price performance page, we showed that market weighting can
be fairly compact. Nonetheless, with so many data points to weight, we need to
spread out.
      So we dedicate an entire worksheet—actually two—to our weighting pro-
cess. The original worksheet is called weightings; the spillover worksheet is called,
cleverly, more weightings. Let’s look at the elements of the weightings worksheet.
Figure 15.7 shows a snippet of our weightings worksheet.
      In columns A through E of the worksheet, we have the elements needed to
begin the weighting process. In order, column A has company names; column B,
tickers; column C, current price; column D, current-year shares outstanding; and
column E, market capitalization, which is simply current price times shares out-
standing. We use rows 22 through 36 in column E to display the individual per-
centages each company represents in the group.
      The market caps of this group, as you can see, are a motley lot. Cisco is
nearly half the market cap, but that was not always the case. Among the compa-
nies that have been knocked back several notches are Nokia and Motorola in
particular. JDSU, which had a $100 billion market cap when it was admitted to
the S&P 500, is now worth about $1 billion. The up-and-comers on a percentage
basis include Juniper (nearly as big now as Motorola, by cap) and most notably
Corning, approaching a 10% weight.
      When we sum the market cap for the coverage group, we see that it is around
$280 billion. At its peak near the century turn, the market cap for this group
likely topped $1 trillion—at a time when the companies were immensely ineffi-
cient and profligate and the technologies they were hawking were wildly over-
hyped. Now that they deliver reliable and truly transformative technology, they
are collectively worth about a quarter of peak value. Stepping off the soapbox, we
brought up that summed or group market cap because that number, residing in
cell E20, is instrumental in every cap-weighted calculation.
      To begin the process, we need two elements for each data point: (1) the
individual company data point itself, and (2) the data point multiplied by its
individual market cap. Let’s take a look at how we market cap–weight P/Es. The
illustration shows historical P/Es from 2003 through 2008 in columns G through
I (note that for presentation purposes we’ve truncated the years 2004 through
Figure 15.7

In this compacted view of a weightings page within the industry matrix workbook, we’ve cut out year 2010 and the interim years between 2003 and 2008. The actual
worksheet sprawls across hundreds of columns. We market-weight so many data points, returns, averages, ratios, and growth rates on the industry matrix workbook that
by necessity we’ve devoted two worksheets to the process.

Weighted Averages
                                                                                                                 Weighted        Weighted Weighted Weighted
                                           Shares Out Market           2003          2008 5_year 2009E            2003            2008      5-year   2009E
                                 Price     2009 (mm) Captlzn            P/E           P/E     P/E      P/E         P/E             P/E       P/E      P/E
ADC Telecom            ADCT         8.65            117   1,013        (49.97)         10.48    38.16   31.82      (50,614)          10,615   38,655   32,227
Adtran                 ADTN       22.66              64   1,440          33.41         14.54   21.06     17.62       48,118         20,938    30,336   25,373
Alcatel                 ALA         3.37         2,296    7,737          (7.53)         17.18    17.18       -    (58,268)         132,890   132,890        -
Big Band               BBND         4.04             64     258                        15.10    15.10   52.84
Ciena                  CIEN        12.29            104   1,284        (14.32)         21.49    21.49        -      (18,380)          27,584       27,584             -
Cisco Systems          CSCO        21.04          6,151 129,422          28.87         17.86   20.84     14.16    3,736,252        2,311,994    2,697,806     1,831,988
Ericsson                ERIC        9.30          3,193 29,695         (14.10)         16.43   15.43    14.33     (418,656)         488,030       458,098       425,614
Extreme Networks       EXTR         2.57            115     296        (41.24)         26.35    35.18   22.01       (12,225)            7,810      10,427         6,523
Corning Inc.            GLW        15.79          1,591 25,120           74.89         11.44   18.48    13.51      1,881,214        287,268       464,278       339,364
JDS Uniphase           JDSU         5.85            228   1,334        (17.94)         25.35 (59.72)    25.98       (23,940)          33,828     (79,689)        34,672
Juniper Networks       JNPR        24.47            556 13,615           90.94         19.68   32.57    28.68      1,238,211        267,900       443,446       390,447
Motorola               MOT          7.03         2,267 15,934            25.23        427.95   111.14        -       401,999       6,819,084    1,770,860             -
Nokia                  NOK         12.95         3,778 48,926            18.43         14.53   15.47    14.53        901,456         710,773      757,007       710,773
Polycomm               PLCM        22.71             87   1,981          35.68         13.82   22.27     17.97        70,685          27,381        44,119       35,610
Sycamore Networks      SCMR         3.08            284     874        (15.81)         76.19     5.71        -       (13,817)         66,587        4,988             -
Tellabs                TLAB         6.59            410   2,705              -         21.88    23.12   22.45               -          59,171      62,522        60,715
                                                        281,635                                                   7,682,036       11,271,852    6,863,327     3,893,306
                                                                                                                                                             (continued)
Figure 15.7 (continued)


Weighted Averages
                                                                                          Weighted   Weighted Weighted Weighted
                                      Shares Out Market    2003   2008   5_year   2009E    2003       2008     5-year   2009E
                              Price   2009 (mm) Captlzn     P/E    P/E    P/E      P/E      P/E        P/E      P/E      P/E
Weight
ADC Telecom           ADCT                          0.4%
Adtran                ADTN                          0.5%
Alcatel                ALA                          2.7%
Ciena                 CIEN                          0.5%
Cisco Systems         CSCO                        46.0%
Ericsson               ERIC                       10.5%
Extreme Networks      EXTR                          0.1%
Corning Inc.           GLW                          8.9%
JDS Uniphase          JDSU                          0.5%
Juniper Networks      JNPR                          4.8%
Motorola              MOT                           5.7%
Nokia                 NOK                          17.4%
Polycomm              PLCM                          0.7%
Sycamore Networks     SCMR                          0.3%
Tellabs               TLAB                          1.0%
                              Simple Average and Market-Weighted Comparisons    •   329



2007 in blank column I). We show the five-year average in column J; and the
2009 P/E in column K. In Columns L through P we have multiplied the indi-
vidual market cap times the P/Es for each of these historical and forward years
and for the five-year period. And, importantly, we have summed the P/E times
cap totals in columns L through P.
       To execute our cap-weighted P/E, we jump back to the company compari-
sons worksheet. To calculate market-weighted P/E for each individual period,
divide the P/E times market cap sum by the sum of total market capitalization.
For 2008, for example, the formula looks like this: Weightings!Q20/
Weightings!E20, where Q20 is the sum of 2008 P/Es times current market caps
and E20 is the summed market cap for the group.
       In the 2008 figure, we immediately see the value of analyzing the group
based on its weighted P/E versus its simple average P/E. The nearly 60 times
simple P/E has been distorted by Motorola’s 400-plus P/E for the year. The more
reasonable P/Es for the industry giants—Cisco, Ericsson, Nokia, and Corning—
show that investors in the group were really on average paying 40 times for pro
forma earnings in 2008. Even that number is high by industry standards, as the
2008 average price for each equity held up relatively better than the earnings
trend, which was decimated by the global economic collapse; 2H08 performance
actually canceled much of 1H08’s EPS.
       Let’s take a careful look at 2009 P/Es, which at first blush seems awfully low.
On an unadjusted basis, deep pro forma losses for Alcatel-Lucent, Ciena, and
Sycamore Networks have dragged down not just the simple average but the
weighted average as well. If we eliminate those negative P/Es on the comparisons
sheet, it will eliminate them from the weighted calculation as well.
       But, because only Alcatel-Lucent among them has much market cap, elimi-
nating those three deeply negative P/Es raises simple and weighted P/Es by a few
percentage points. The 2010 market-weighted group average P/E is likely a truer
picture of how much investors are willing to pay for the group. At about 15 times,
the group—with low debt, lots of cash, few pension obligations, broad global
exposure, and positioning squarely in a secular growth area—trades at a slight
discount to the market.
       The Big Five comparable historical valuations—price to earnings, sales,
book value, cash flow, and relative P/E—look much the same as the other indus-
try valuation grinds. Each has historical data for the individual companies for at
least the preceding five years; forward P/Es for the next two years for all members
of the group; simple and weighted group averages for each years individual com-
pany five-year averages; and the simple average of the individual five-year
averages.
330   •   Relational Valuation: The Industry Matrix Workbook and Peer Derived Value



       What’s different is that we’ve taken time to calculate a weighted average of the
five-year average for these five valuation measures. In this entire worksheet, which
runs nearly 500 lines and includes about 20 data groupings, we’ve only done this
last step for the Big Five historical comparables.
       We’ve done so because we need this information for the final piece of our
valuation puzzle, called Peer Derived Value.
                                               Chapter 16
           PEER DERIVED VALUE




Confronting the Peer Value Problem
If you read enough research reports, every so often you’ll come across a sentence
like this: “The XYZ shares, which typically trade at a premium to the peer group,
now trade at a discount.” Okay, now what?
       Intuitively, we know this observation has captured a useful discrepancy and
perhaps (emphasis mine, doubled) represents a value opportunity. But is it an
investable discrepancy? If we don’t act on that observation within one week, is it
still valid? What caused the discrepancy, by the way: Has the peer group P/E
moved? Or was it the asset itself? More to the point, is this value discrepancy in
and of itself sufficient justification to act?
       The fact is that this kind of observation amounts to a tip or a hunch, and
one with a very uncertain shelf life. Yet if we have a means to quantify this obser-
vation and to do so in real time, it can provide powerful information that informs
valuation analysis and the asset decision.
       Most investors don’t have access to useful tools for quantifying this obser-
vation or for trading on it. In a world awash in financial volumes on comparable
historicals, DFCF, and DDM, you would be hard pressed to unearth a volume
devoted to deriving or adjusting asset value based on the peer group relationship.

                                                                                 331
332   •   Relational Valuation: The Industry Matrix Workbook and Peer Derived Value



Many analysts and investors seeking valuation information from these kinds of
relationships rely on the familiar or, conversely, on exotic formulations they con-
coct in their own labs.
       The most common methodology for identifying valuation discrepancies is
to use relative P/Es. The problem with relative P/Es, of course, is they encompass
not just the peer group but the whole market. To tortuously paraphrase Groucho,
we wouldn’t want to join a club insufficiently indiscriminate to include us along
with everybody else.
       So why not replicate the relative P/E process but on a smaller and more
focused scale? It can be done, but it is a far from simple process. Implicit in rela-
tive P/Es, remember, is a benchmark—the S&P 500—as well as benchmark earn-
ings. Industry indexes abound, but none is likely a perfect match for your exact
coverage. And even if it were, the most we could get is industry index earnings;
we regard earnings as insufficient for the valuation decision.
       How about building your own index? It would be simple enough to build a
weighted earnings scale for a given peer group; we pretty much weight everything
we come across. But we would also need to build a cap-weighted index, and that
is far from simple. We need to declare a base-line value and beginning point in
time, determine a divisor, accommodate member changes, and deal with a range
of other issues. Moreover, operating in a vacuum, we would lack the validation
of the market and investing public; we might never know if our baseline or divi-
sor assumptions were correct.
       After considering these limitations and challenges, we concluded we needed
to build our own methodology. Fortunately, all the groundwork we’ve laid so far
in the individual company workbooks and the industry matrix steers us almost
directly to the solution. The wealth of data and links to real-time pricing enable
us to create the solution and keep it current efficiently and painlessly.
       As the previous paragraph implies, if you’ve skipped ahead in hopes of snar-
ing the latest valuation magic bullet, we’ve got bad news. The essential inputs in
this process are contained—embedded, really—in all the prior work. That is not
to say publicly available information cannot be used to make this technique work.
But we would argue that to attain maximum flexibility and veracity in the pro-
cess, it is best to control the inputs.


The Elements of PDV
Let’s revisit the earlier statement that triggered this exercise. “The XYZ shares,
which typically trade at a premium to the peer group, now trade at a discount.”
                                                           Peer Derived Value   •   333



If we break it down, we see that the subject and dependent clause in this sentence
contain an historical observation: XYZ typically trades at a premium to the peer
group. There is also the current circumstance that captures the value discrep-
ancy: XYZ now trades at a discount to the peer group.
       There are a few other things implicit in such a casual observation. Though
not stated, the value discrepancy has likely been identified based on price to earn-
ings, the market’s most common and familiar valuation measure. We know that
actual stock movement is sometimes contrary to P/E trends; in the early days of
the 2009 recovery, high P/E stocks hugely outperformed low P/E stocks because
investors preferred to buy future, not present, earnings. In designing our peer
group valuation process, we did not want to rely solely on P/E. We knew that if
we wanted to flesh out this suggested value discrepancy, we would have at a mini-
mum five comparable historical metrics to pull from the value kit. The Big Five
historical comparables used almost universally are price to earnings, price to
sales, price to book value, price to cash flow, and relative P/E
       The historical relationship between an equity and its peer group, the first
part of this puzzle, is static. Both the prices at which the peer group stocks traded
and the financial statement data (e.g., earnings, revenues, and book value)
included in the historical comparables are frozen in time. In our industry matrix,
we calculated five-year average P/Es, price to sales, and other comparables for
each company. We did the same for the peer group. And, for this subset of Big
Five historical comparables, we also calculated a weighted five-year average, in
order to minimize the impact of fast-growing or troubled outliers.
       Armed with that knowledge, it is straightforward to calculate and quantify
the relationship between a company’s five-year average price-based metric and
the peer group’s five-year average weighted price-based metric. For simplicity’s
sake in the following example, we’ll let P/E stand in for each of the Big Five.
       Now that we’ve quantified the historical relationship of each individual P/E
to the group’s weighted average, we need to quantify the relationship between the
current (i.e., two-year forward) P/E for each company with that of group current
(i.e., two-year forward) weighted average. That is a slightly more elusive set of
circumstances, in that the values and thus the value relationships are moving
around based on changes in the inputs (i.e., sales and earnings) and in the asset
prices. Because future events can’t be pinned to the board like butterfly wings but
are ever shifting, this set of relationship needs to be informed by real-time pricing
and real-time adjustments to the individual company inputs.
       Fortunately, we link our peer group valuation model to the industry matrix,
which in turn is linked to the individual workbooks. We also have real-time pric-
ing on both the individual company models and within the industry matrix
334   •   Relational Valuation: The Industry Matrix Workbook and Peer Derived Value



workbook. Thus our peer group calculation can capture these adjustments in real
time. While we’ve been using the term P/E in the above discussion, we know the
peer group relative value assessment needs to be comprehensive. So we need to
scale a model that adds valuation color based on revenue, book value, cash flow,
and relative P/E.
      We have identified the specific challenge. Basically, we need to quantify the
historical relationship of the asset to its peer group, quantify the current relation-
ship (which necessarily has some forward elements) of the asset to its peer group,
quantify the relationship between the historical relationship and the current rela-
tionship, and adjust the current price accordingly.
      Note that while we seek to quantify the relationship between relationships,
we are indifferent to the actual content of any valuation in and of itself. In other
words, the fact that one stock traded at a P/E of 22 times for the preceding five
years and another traded at 11 times has no bearing on our PDV value calcula-
tion. In using Peer Derived Value, all we care about is the degree to which the first
stock currently varies from its 22 times historical P/E and the second stock cur-
rently varies from its 11 times historical P/E.
      All that remains is to erect the infrastructure required to establish this set
of relationships. While this set of relationships may seem esoteric, I am not exag-
gerating when I say the required elements are staring us in the face.


Physical Structure of the PDV Worksheet
Historical Variation versus Weighted Average
For convenience, our PDV calculations occur on a worksheet within the industry
matrix workbook. Later, the individual company conclusions will be linked from
the industry matrix workbook back to the individual company workbooks to
influence valuation and the asset decision.
      Take a look at the PDV worksheet represented in truncated form in Figure
16.1. The area at top left shows our conclusions. The area at the right shows the
process; and the area below shows the calculations that support our processes.
So, working forward rather than backward, it makes sense to begin in the lower
area, approximately around row 22.
      Rather than start with P/Es, as the market always seems to, we will first look
at price to sales. But before we begin, we want to be sure to incorporate the most
up-to-date comparable historical ratio information. Accordingly, we want to
open each individual workbook and refresh the price, which in turn refreshes the
valuation reading. We also want to refresh prices on the industry matrix itself.
Figure 16.1

Peer Derived Value (PDV) is an original and proprietary valuation methodology that seeks to value an equity based on current variation from its historical relation to a
user-specified peer group. The process, which sprawls across and particularly down its worksheet, is necessarily compacted in Figure 16.1.

Peer Derived Value (PCV)
                                                            Premium/
                                                             Dscnt to                                                                                      Calculated
                                                 Current      current                                                                                         Value
Companies                            PDV          Price        price                  P/S         P/BV          P/CF          P/E        Rel P/E           (Average)
ADC Telecom                      12.68         8.65         1.47                 14.56         5.59         22.90         9.61         10.76               12.68
Adtran                           20.83         22.66        0.92                 21.58         15.45        23.11         20.80        23.20               20.83
Alcatel-Lucent                   4.20          3.37         1.25                 6.72          2.53         3.35          -            -                   4.20
BigBand                          4.93          4.04         1.22                 5.03          5.19         8.14          1.37         -                   4.93
Ciena                            7.89          12.29        0.64                 22.83         8.72         -             -            -                   7.89
Cisco Systems                    22.77         21.04        1.08                 22.80         20.96        25.63         21.74        22.74               22.77
Ericsson                         9.52          9.30         1.02                 12.46         9.57         8.97          8.03         8.57                9.52
Extreme Networks                 5.21          2.57         2.03                 4.72          2.55         5.85          3.68         9.26                5.21
Corning Inc.                     17.30         15.79        1.10                 13.43         18.24        20.23         16.43        18.20               17.30
JDS Uniphase                     4.79          5.85         0.82                 18.28         6.34         (8.57)        -            7.90                4.79
Juniper Networks                 19.97         24.47        0.82                 25.04         13.87        14.28         22.45        24.20               19.97
Motorola                         7.55          7.03         1.07                 7.90          6.58         2.98          11.51        8.78                7.55
Nokia                            13.64         12.95        1.05                 18.73         14.74        13.26         12.22        9.25                13.64
Polycomm                         23.56         22.71        1.04                 24.13         15.85        29.65         23.02        25.16               23.56
Sycamore Networks                1.20          3.08         0.39                 2.00          2.11         -             0.28         1.63                1.20
Tellabs                          5.63          6.59         0.85                 6.91          3.93         6.14          5.34         5.83                5.63
Average                                                     1.05
                                                                                                                                                            (continued)
Figure 16.1 (continued)


Peer Derived Value (PCV)

                                                                                                                    Preemium/
                                                                                                                    Discount of
                                                     Historical                                                 Historical Variation
                                5-Year Avg          Variation vs.          Frwrd (09-10)      Current Variation     to Forward
                                 Multiple            Wghtd Avg               Multiple          vs. Wghtd Avg         Variation         Price
Companies                  Price to Sales
ADC Telecom                                  1.83               0.48                   0.74                 0.29                1.68            8.65
Adtran                                       3.71               0.98                   2.67                 1.02                0.95           22.66
Alcatel-Lucent                               0.99               0.26                   0.34                 0.13                2.00            3.37

Tellabs                                      2.32                   0.61               1.52                 0.58                1.05            6.59
Simple Average                               3.82                                      2.55                                     1.40
Weighted Average                             3.80                                      2.60
                                                                                      -32%

Companies                  Price to Book Value
ADC Telecom                               2.58                  0.59                   2.25                 0.91                0.65            8.65
Adtran                                    4.06                  0.93                   3.35                 1.36                0.68           22.66
Alcatel-Lucent                            2.01                  0.46                   1.51                 0.61                0.75            3.37

Tellabs                                      1.48               0.34                   1.39                 0.57                0.60            6.59
Simple Average                               3.27                                      1.99                                     0.87
Weighted Average                             4.38                                      2.46
                                                                                      -44%
      Figure 16.1 (continued)


      Peer Derived Value (PCV)

                                                                                                                          Preemium/
                                                                                                                          Discount of
                                                          Historical                                                  Historical Variation
                                      5-Year Avg         Variation vs.          Frwrd (09-10)       Current Variation     to Forward
                                       Multiple           Wghtd Avg               Multiple           vs. Wghtd Avg         Variation         Price
      Companies                  Price to Cash Flow
      ADC Telecom                                31.22               2.05                  10.01                  0.77                2.65            8.65
      Adtran                                    18.24                1.20                  15.17                   1.17               1.02           22.66
      Alcatel-Lucent                             13.61               0.89                  11.62                  0.90                0.99            3.37

      Tellabs                                   13.00                0.85                   11.83                 0.92                0.93            6.59
      Simple Average                            15.00                                      14.23                                      0.95
      Weighted Average                          15.24                                      12.93
                                                                                            -15%

      Companies                  Price to Earnings
      ADC Telecom                               38.16                1.92                  24.76                  1.73                1.11            8.65
      Adtran                                    21.06                1.06                  16.53                  1.15                0.92           22.66
      Alcatel-Lucent                             17.18               0.86                                                                             3.37

      Tellabs                                    23.12                   1.16              20.54                  1.43                0.81            6.59
      Simple Average                            15.39                                      16.81                                      0.87
      Weighted Average                          19.89                                      14.33
                                                                                           -28%
337
Figure 16.1 (continued)


Peer Derived Value (PCV)

                                                                                                                    Preemium/
                                                                                                                    Discount of
                                                    Historical                                                  Historical Variation
                               5-Year Avg          Variation vs.          Frwrd (09-10)       Current Variation     to Forward
                                 Multiple           Wghtd Avg               Multiple           vs. Wghtd Avg         Variation         Price
Companies                  Relative P/E
ADC Telecom                                 2.26                   2.13               1.35                  1.71                1.24            8.65
Adtran                                      1.28                   1.21               0.93                  1.18                1.02           22.66
Alcatel-Lucent                                                                                                                                  3.37

Tellabs                                     1.39                   1.31                1.17                 1.48                0.88            6.59
Simple Average                              1.65                                      0.99                                      1.16
Weighted Average                            1.06                                      0.79
                                                                    Peer Derived Value     •   339



      We begin, as we always seem to, with a list of our covered companies in
column A. In Figure 16.2, we focus on the first part of the equation: the indi-
vidual equities’ variation from the historical weighted average for price to sales.
In column C, for each company we link to that company’s five-year average price/
sales ratio; these are available on the company comparisons worksheet within the
industry matrix. In our example, we are showing 16 companies within the col-
umn. In the seventeenth row in this array, we link to the simple average of all
companies’ five-year price to sales, and in the eighteenth row we link to the
weighted average of all companies’ five-year weighted average. We see that the
five-year average P/S is 3.82.
      In other words, for the preceding five years, you could buy this group on
average for just less than four times annual revenue. Given that there are few
“specials” and surprises in revenue, the weighted average is remarkably close at
3.80. The other valuation measures will be loaded with (mainly unhappy) sur-



Figure 16.2

Communications equipment companies traded at an average four times revenue during 2004–2008.
In column D, we quantify each company’s variation from that average over the time period.

Peer Derived Value (PDV)
                                               5-Year Avg           Historical Variation
                                                Multiple              vs. Wghtd Avg

Companies                                     Price to Sales
ADC Telecom                                                 1.83                    0.48
ADTRAN                                                      3.71                    0.98
Alcatel                                                     0.99                    0.26
BigBand                                                     3.28                    0.86
Ciena                                                       4.67                    1.23
Cisco Systems                                               5.07                    1.33
Ericsson                                                    1.94                    0.51
Extreme Networks                                             1.71                   0.45
Corning Inc.                                                5.56                    1.46
JDS Uniphase                                                4.66                    1.22
Juniper Networks                                            6.08                    1.60
Motorola                                                     1.15                   0.30
Nokia                                                       1.68                    0.44
Polycomm                                                    3.01                    0.79
Sycamore Networks                                          13.48                    3.54
Tellabs                                                     2.32                    0.61
Simple Average                                              3.82
Weighted Average                                            3.80
340   •   Relational Valuation: The Industry Matrix Workbook and Peer Derived Value



prises and specials; for that reason in PDV we will always measure variation off
the peer group on a cap-weighted basis.
      In column D, for each stock we calculate the variation to the historical
group-weighted average P/S. Assuming group-weighted average historical P/S is
in cell C45, beginning in cell D28 and on through cell D43, we divide each com-
pany’s five-year average P/S by the weighted average.
      Let’s take a look at a few examples to see what the data—and the data rela-
tions—tell us. The first two stocks in the column are ADC Telecom, principally
a provider of connectivity solutions for carrier, enterprise, and cable networks;
and ADTRAN, a technology fast-follower and price leader for carrier and enter-
prise customers. We did not choose these two because they are next to one
another (nor for the near-fantastic coincidence that both have a CFO named
James Mat(t)hews). We chose them because one has had a stable last few years
and the other has effectively reinvented itself and taken some bumps in the pro-
cess. Presumably, the market will have weighed in on these very different
experiences.
      Without delving too deeply into each company’s situation, we note that
investors have always liked ADTRAN—and evidently like it now a bit more—
because it has an unparalleled customer base of top-tier telecommunications
companies (telcos), executes well when it attacks new markets such as enterprise
networking, and (here’s the “like it more now” part) a pricing strategy in which
new products are priced at discounts to the market and new iterations of older
products go out at successively lower price points.
      ADC, under CEO Robert Switz, has in our view followed a courageous path
to forge itself anew. ADC used to serve lots of traditional telco equipment niches,
though not particularly well; rather than remain a second-tier player with an
unfocused strategy, the company repositioned itself as a connectivity and access
player. Along the way, some missteps occurred; investors were not happy with a
failed all-stock bid for a rival. More specifically to the price/sales equation, inves-
tors may perceive the company as now playing in lower-margined niches and thus
are less willing to pay what they once did for revenues.
      ADC’s five-year historical price/sales ratio of 1.83 is less than half the group
average; specifically, ADCT traded at 48% of the group’s weighted average five-
year P/S. ADTRAN’s historical P/S was right around the average at 3.71; that is
98% of the five-year weighted peer group price/sales ratio. Before we delve into
the reasons for the market’s very different treatment of these industry peers, let’s
see how these relationships to the peer group hold up in the current
environment.
                                                            Peer Derived Value   •   341



Current Variation versus Weighted Average
In column E of Figure 16.3, we link from the company comparisons worksheet
within the industry matrix to get the forward (2009 and 2010) P/S ratios for our
16 companies. Because we need an average in this column, our formula is an
average of the two links. For ADC Telecom, the formula would be something like
  (‘Cmprsn Tlcm’!N410 ’Cmprsn Tlcm’!O410)/2, where cell N410 on Cmprsn
Tlcm is the 2009 calculated price/sales (based on our income statement modeling
and linked originally from the ADCT individual workbook) and cell O410 is the
2010 value. We proceed down line with the two-year forward average P/S for each
company. In cell E44, we link to the simple average in this series, which is 2.55;
in cell E45, we link to the weighted average in this series, which is 2.60.
      Figure 16.3 shows the “new reality,” or at least the reality as it stood in mid-
summer 2009. The entire market has been through a very rough patch. For our
group, whereas investors were once willing to pay nearly four times sales for the
group, they are now only ponying up 2.6 times sales—a 32% discount to the five-
year average.
      In column F, we replicate the value relationship calculations we used in
column D, but instead of determining for each stock the variation to the historical
group-weighted average P/S, for each stock we want to calculate the variation to
the forward group-weighted average P/S.
      How are our sample companies holding up in this brave new world? Spe-
cifically, how are their value relationships to the peer group faring? ADC Tele-
com, we learn, is trading at a two-year average P/S of just 0.74, or just under 30%
of the group average. ADTRAN, by contrast, has retained its value relationship
to the group. Whereas ADTRAN traded at 98% of the group average for the pre-
ceding five years, its current (forward) two-year average P/S ratio is 2.67, or 102%
of the group average.
      In other words, despite the gyrations in the market, investors have more or
less maintained ADTRAN’s historical relationship to the peer group, at least on
a price/sales basis. ADC has not been so lucky. Despite beginning at a substantial
discount to the peer group on P/S, ADC’s discount has increased to the point at
which its P/S—once half the group average—is now less than one-third.


Premium/Discount of Historical Variation to Forward Variation
While ADC has not fared well with investors, perhaps investors are misreading
the stock and a value discrepancy has been created. In column G of Figure 16.4,
342   •   Relational Valuation: The Industry Matrix Workbook and Peer Derived Value



Figure 16.3

The economic recession and stock market collapse may have wracked some sectors harder, but
technology was not spared; the group at midsummer 2009 traded at just 2.6 times sales or at a 32%
discount to its preceding five-year average.


Peer Derived Value (PDV)
                                                     Historical       Frwrd          Current
                                  5-Year Avg        Variation vs.    (09-10)       Variation vs.
                                   Multiple          Wghtd Avg       Multiple       Wghtd Avg

Companies                         Price/ Sales
ADC Telecom                                 1.83              0.48          0.74             0.29
ADTRAN                                      3.71              0.98          2.67             1.02
Alcatel                                     0.99              0.26          0.34             0.13
BigBand                                     3.28              0.86          1.80             0.69
Ciena                                       4.67              1.23          1.72             0.66
Cisco Systems                               5.07              1.33          3.20             1.23
Ericsson                                    1.94              0.51          0.99             0.38
Extreme Networks                             1.71             0.45          0.64             0.24
Corning Inc.                                5.56              1.46          4.48             1.72
JDS Uniphase                                4.66              1.22          1.02             0.39
Juniper Networks                            6.08              1.60          4.07             1.56
Motorola                                     1.15             0.30          0.70             0.27
Nokia                                       1.68              0.44          0.79             0.31
Polycomm                                    3.01              0.79          1.94             0.75
Sycamore Networks                          13.48              3.54         14.23             5.47
Tellabs                                     2.32              0.61          1.52             0.58
Simple Average                              3.82                            2.55
Weighted Average                            3.80                            2.60
                                                                           -32%


we seek to quantify the relationship of the historical to the forward P/S with a
hope of discovering such discrepancies. Figure 16.4 quantifies the change in the
historical peer group relationship to the current peer group relationship.
      In column G and beginning in cell G28 in our example, for each company
we divide the historical variation to the weighted average, by the current variation
to the weighted average. In ADC’s case, its 0.48 historical variation divided by its
0.29 current (forward) variation results in a premium of 1.68. A value below 1 in
this column signals that, on this measure alone, a stock is overvalued relative to
its peers. A value above 1 suggests that the stock is undervalued relative to
peers.
      In other words, based solely on the variation from its historical norm in
price/ sales, ADC seems significantly undervalued. ADTRAN sends a different
Figure 16.4

Our peer-derived calculation suggests that, on a price/sales basis, ADTRAN has very much maintained its historical relationship with its peers, while ADC Telecom
trades at a discount to its normal spot in the pecking order.

Peer Derived Value (PDV)
                                                             Historical                                Current        Premium/Discount of               Peer
                                          5-Year Avg        Variation vs.      Frwrd (09-10)         Variation vs.    Historical Variation to      Derived Value
                                           Multiple          Wghtd Avg           Multiple             Wghtd Avg         Forward Variation              (PDV)

Companies                                 Price/Sales
ADC Telecom                                         1.83               0.48                   0.74             0.29                        1.68              8.65
ADTRAN                                              3.71               0.98                   2.67             1.02                        0.95             22.66
Alcatel                                             0.99               0.26                   0.34             0.13                        2.00              3.37
BigBand                                             3.28               0.86                   1.80             0.69                        1.24              4.04
Ciena                                               4.67               1.23                   1.72             0.66                        1.86             12.29
Cisco Systems                                       5.07               1.33                   3.20             1.23                        1.08             21.04
Ericsson                                            1.94               0.51                   0.99             0.38                        1.34              9.30
Extreme Networks                                     1.71              0.45                   0.64             0.24                        1.84              2.57
Corning Inc.                                        5.56               1.46                   4.48             1.72                        0.85             15.79
JDS Uniphase                                        4.66               1.22                   1.02             0.39                         3.12             5.85
Juniper Networks                                    6.08               1.60                   4.07             1.56                        1.02             24.47
Motorola                                             1.15              0.30                   0.70             0.27                         1.12             7.03
Nokia                                               1.68               0.44                   0.79             0.31                        1.45             12.95
Polycomm                                            3.01               0.79                   1.94             0.75                        1.06             22.71
Sycamore Networks                                  13.48               3.54                  14.23             5.47                        0.65              3.08
Tellabs                                             2.32               0.61                   1.52             0.58                        1.05              6.59
Simple Average                                      3.82                                      2.55                                         1.40
Weighted Average                                    3.80                                      2.60
                                                                                             -32%
344   •   Relational Valuation: The Industry Matrix Workbook and Peer Derived Value



signal. After trading at a five-year P/S fractionally below the weighted group aver-
age, it currently trades at a P/S fractionally above the group average. Its discount
of historical variation to forward variation is 0.95, suggesting on this metric alone
that it is a tad, or half a tad, overvalued.
      In any event, the PDV calculation appears to have uncovered a value dis-
crepancy for ADC based on price/sales, while suggesting that ADTRAN’s value
relationship with the peer group is largely unchanged.


Replicating the Process
By this point, you’ve learned that we want to apply as much information as pos-
sible to inform the value decision. That’s because every valuation metric has a
“yeah, but” element. We analyze P/Es, but with the caveat that P/Es can some-
times be counterintuitive, given investors’ anticipatory impulses and desire to
purchase future earnings rather than current earnings. Price/sales has informa-
tion value, but companies can sometimes sacrifice margin to maintain or expand
sales. Price/book value has information value, but we’ve discussed the way man-
dated impairments and aggressive business development can distort stockhold-
ers’ equity. Price/cash flow has value, but depreciation schedules can sometimes
mislead on cash progress.
      We have calculated a wealth of individual asset and peer group valuation
analysis, and we need to bring all of it to bear to minimize any distortions from
any single valuation technique. Once we’ve set up the grid for price/sales, we can
replicate the process for price/earnings, price/book, price/cash flow, and even
relative P/E. (I once saw a sign in a beach house: “Relatives of relatives may not
bring relatives”—but this is not a beach house.)
      It is in fact simple to replicate the process for those other inputs. Let’s move
through the other price-based historical comparables. Along the way, we’ll check
in with ADC Telecom and ADTRAN to see how their historical value relation-
ships are holding up.


PDV Price/Book
The grid to determine PDV price/book is identical to the grid we set up for price/
sales PDV. The only difference is that we import and link to the price/book value
information on our company comparisons worksheet rather than P/S data. In
Figure 16.5, the companies are in column A. Historical price/book value is in
column C; the historical variation to the weighted group average is in column D.
The forward two-year average P/BV multiple for each company and for the group
is in column E. The current variation to the weighted group average is in column
                                                         Peer Derived Value   •   345



F. And the premium or discount of historical variation to forward variation is in
column G.
       With this metric, we begin to see the value of using weighted group aver-
age. The very low price/book ratios of a couple of outliers (Sycamore Networks
and Tellabs) distorts the simple average, which is 3.27 on an historical basis; on
a weighted basis, the average is 4.38. Similarly, for the forward P/BV multiple,
the weighted average of 2.46 better represents the group in total than the simple
average of 1.99. The difference between historical weighted group P/BV and
current weighted group P/BV is –44%, a meaningfully larger decline than the
–32% decline we recorded between historical weighted group P/S and current
weighted group P/S. Why the discrepancy? Not a dollar of revenue has been
impaired, but lots of stockholders’ equity (the underlying value in book value)
has been impaired under FAS 142. So the 44% decline represents not just the
market price decline but also the smaller collective pool of equity against which
price is assessed.
       As for our sample companies, ADC Telecom has a five-year average P/BV
of 2.58, representing 0.59 of the group weighed P/BV. Looking forward, ADC’s
two-year forward price/book is 2.25; that represents 0.91 of the forward group-
weighted P/BV of 2.46. That would seem to suggest investors have bid up the
stock. How does that jibe with the P/S experience, which seemed to imply the
opposite?
       Let’s look under the covers; specifically, we go to the R&V worksheet within
our ADC Telecom individual workbook to examine the balance sheet. We see
that the five-year average dollar value of stockholders’ equity for ADC from 2004
through 2008 is $846 million. But the two-year average of actual (2009) and
forecast (2010) stockholders’ equity is $377 million. In other words, more than
all of the increase in investors’ seeming willingness to pay up for the ADCT stock
is explained not by price movement, but by retained-earnings decline flowing
from asset impairment. This is a stark example of why we do not rely on any one
input in determining PDV.
       As for ADTRAN, its five-year average P/BV of 4.06 captures 0.91 of the
group-weighted average. Looking forward, its current (two-year forward) P/BV
of 3.35 has declined, though not as much as the weighted group P/BV has declined.
Accordingly, ADTRAN’s current variation versus the weighted average is 1.36.
Within the G column, where we derive the premium or discount of historical
variation to forward variation, ADTRAN scores a 0.68. To reiterate, any value
below 1 in this column signals that a stock is overvalued relative to its peers. A
value above 1 suggests that the stock is undervalued relative to peers.
       On this measure alone, ADTRAN and ADC appear overvalued relative to
their peers.
Figure 16.5

On a price/book basis, ADTRAN and ADC Telecom each appear to trade at a similar discount to their normal relationship to the peer group.

Peer Derived Value (PDV)
                                                       Historical                                                 Premium/Discount of              Peer
                                   5-Year Avg         Variation vs.       Frwrd (09-10)       Current Variation   Historical Variation to     Derived Value
                                    Multiple           Wghtd Avg            Multiple           vs. Wghtd Avg        Forward Variation             (PDV)

                                  Price to Book
Companies                             Value
ADC Telecom                                   2.58                0.59                2.25                 0.91                       0.65               8.65
Adtran                                        4.06                0.93                3.35                 1.36                       0.68              22.66
Alcatel                                       2.01                0.46                 1.51                0.61                       0.75               3.37
BigBand                                        4.15               0.95                1.82                 0.74                       1.29               4.04
Ciena                                         2.85                0.65                2.26                 0.92                       0.71              12.29
Cisco Systems                                 5.31                1.21                2.99                 1.22                       1.00              21.04
Ericsson                                      2.90                0.66                1.59                 0.64                       1.03               9.30
Extreme Networks                              2.64                0.60                1.49                 0.61                       0.99               2.57
Corning Inc.                                  3.97                0.91                1.93                 0.78                        1.15             15.79
JDS Uniphase                                  2.68                0.61                1.39                 0.56                       1.08               5.85
Juniper Networks                              2.22                0.51                2.20                 0.89                       0.57              24.47
Motorola                                      2.87                0.66                1.73                 0.70                       0.94               7.03
Nokia                                         5.00                 1.14               2.47                 1.00                        1.14             12.95
Polycomm                                      2.27                0.52                1.82                 0.74                       0.70              22.71
Sycamore Networks                             1.08                0.25                0.89                 0.36                       0.69               3.08
Tellabs                                       1.48                0.34                1.39                 0.57                       0.60               6.59
Simple Average                                3.27                                    1.99                                            0.87
Weighted Average                              4.38                                    2.46
                                                                                     -44%
                                                           Peer Derived Value   •   347



PDV Price/Cash Flow
Again, for price/cash flow (P/CF) we replicate the PDV grid in columns A through
H. Here for the first time we have to deal with outliers with values that are so
severely distorting that we need to exclude them from the grid. Both Ciena and
Sycamore had negative cash flows for most of the historical comparison period.
These (necessarily subjective) exclusions prevent our PDV system from function-
ing like a true closed loop.
      We see in Figure 16.6 that the historical (five-year) peer group–weighted P/
CF ratio is 15.24. The forward peer group–weighted P/CF is 12.93, about 15%
below the historical peer-weighted P/CF. How do we square that with the much
larger declines of 32% for P/S and 44% for P/BV? Simply, the group has begun to
earn better; that is supplementing the steady contribution to cash flow from
depreciation and amortization and is driving down P/CF ratios more quickly
than the market is driving down prices.
      ADC Telecom, with weak earnings in the historical period, has a historical
P/CF ratio of 31.22, slightly more than double the historical weighted peer group
average P/CF ratio of 15.24. As its earnings have recovered, its forward P/CF ratio
has been cut to one third that level, or to about 10 times; its current variation to
the current weighted average is 0.77. It then follows that its relation of historical
variation (2.05) to forward variation (0.77) results in a premium of 2.65. Given
that a value above 1 suggests that the stock is undervalued relative to peers, on
this measure ADC appears to represent good value.
      Well-behaved ADTRAN has a historical P/CF of 18.24, representing 1.20 of
the group historical weighted mean. ADTRAN’s forward P/CF multiple of 15.17
represents 1.17 of the group forward mean. ADTRAN’s premium of historical
variation to forward variation is 1.02, suggesting that on P/CF the company is
trading right around where it always has in relation to the peer group; there is no
value discrepancy here.


PDV Price/Earnings and Relative P/E
Let’s construct the now-familiar PDV grid a fourth time, to accommodate PDV
price/ earnings. We’ve actually combined both P/E and relative P/E in Figure 16.7.
In column C, we see that historical peer group–weighted P/E is 19.9 times. Net
income in the technology space is a fungible number, we’ve learned. Our P/Es are
based on pro forma earnings, consistent with the consensus practice for this
technology niche. For our five-year time frame, the 20-times multiple really bor-
rows from two eras: (1) the first few years (2004 to 2006), when technology was
Figure 16.6

The decline in the group’s price/cash flow from historical levels to the present has been much less than similar declines in P/S and P/BV. We think that is because
these companies now have more predictable and stronger earnings.

Peer Derived Value (PDV)
                                                                                                                                           Preemium/Discount
                                                                                                                                               of Historical
                                               5-Year Avg            Historical Variation      Frwrd (09-10)         Current Variation vs. Variation to Forward
                                                Multiple               vs. Wghtd Avg             Multiple                Wghtd Avg               Variation

Companies                                 Price to Cash Flow
ADC Telecom                                               31.22                        2.05                  10.01                    0.77                     2.65
Adtran                                                    18.24                        1.20                  15.17                     1.17                    1.02
Alcatel                                                   13.61                        0.89                  11.62                    0.90                     0.99
BigBand                                                   26.63                        1.75                  11.21                    0.87                     2.02
Ciena                                                                                      -                22.04                     1.70                        -
Cisco Systems                                                17.39                      1.14                 12.11                    0.94                     1.22
Ericsson                                                     12.27                     0.80                 10.80                     0.83                     0.96
Extreme Networks                                             27.41                     1.80                  10.22                    0.79                     2.27
Corning Inc.                                                 14.50                     0.95                   9.61                    0.74                     1.28
JDS Uniphase                                                (8.01)                   (0.53)                   4.64                    0.36                   (1.46)
Juniper Networks                                            20.70                      1.36                  30.11                    2.33                     0.58
Motorola                                                     10.04                     0.66                  20.12                    1.56                     0.42
Nokia                                                        13.27                     0.87                  11.00                    0.85                     1.02
Polycomm                                                    26.30                      1.73                  17.09                    1.32                     1.31
Sycamore Networks                                                -                         -                 20.17                    1.56                        -
Tellabs                                                      13.00                     0.85                  11.83                    0.92                     0.93
Simple Average                                              15.00                                           14.23                                             0.95
Weighted Average                                            15.24                                           12.93
                                                                                                            -15%
                                                          Peer Derived Value   •   349



still being accorded giddy multiples even as the trend was receding and (2) the
latter stage (2007 to 2008), when technology—cash rich, low debt, lacking pen-
sion obligations, and with globally dispersed markets—was emerging as one of
the more stable and thus market-mimicking niches.
       Even within this increasingly positive climate, not every company could
exploit the favorable trends in this space. For the five-year period, we had to dis-
card JDSU, which was mainly producing losses. We also excluded Motorola’s
2008 P/E, which was in the midhundreds.
       Among our two test cases, ADC—in the midst of completely reinventing
itself in the 2004–2008 period—was predictably earnings starved. Its historical
P/E of 38.16 times equates to 1.92 in relation to the historical peer group–weighted
P/E. Stable ADTRAN, adding new assets at this time (e.g., NetVanta) but not
really disturbing its equilibrium, traded at a 21.06 P/E, or at 1.06 of the peer-
weighted average.
       The forward peer group average weighted P/E drops to 14.33, actually below
the market P/E when this calculation was made. The submarket P/E for this niche
reflects sound balance sheets and relatively consistent earnings at a time when
earnings in other sectors were falling apart. Investors in the current period have
again made technology the largest weighting in the S&P 500. While investors are
plainly drawn to technology’s stability, its growth in overall S&P 500 weighting
is equally a function of the sectors—financial and consumer discretionary,
mainly—from which investors have recoiled.
       ADC Telecom’s 24.8 P/E for the forward period amounts to 1.73 of the for-
ward peer group average P/E. ADTRAN’s l6.5 times P/E is 1.15 times the peer
average. ADC has a 1.11 premium of historical variation to forward variation;
ADTRAN has a 0.92 discount of historical variation to forward variation. ADC
is just a little bit cheap compared to its normal P/E relationship to the group, and
ADTRAN just a bit pricey.
       Our final PDV grid, relative P/E, also requires some adjustments. We have
had to eliminate the negative forward inputs from Alcatel-Lucent and Ciena; we
have had to exclude a few inputs from Motorola’s historical and forward years.
Mainly because of very low relative P/Es for large-cap foreign names (Nokia and
Ericsson) along with Cisco’s low relative P/Es, the historical peer group–weighted
relative P/E is 1.06, or barely more than the market’s. The forward period is even
more intriguing; sub-1 relative P/Es are rife in the group, and the forward peer
group–weighted relative P/E is 0.79.
       ADC had twice the relative P/E in the historical period, at 2.13, and it has
less than two times the group average for the forward period, at 1.71. By now we
know that this will result in a plus-1 premium of historical variation to forward
Figure 16.7

Among our sample companies, ADC was starved for earnings in the preceding five years, resulting in high historical P/Es; as earnings have risen, its P/Es have come
down, suggesting a possible value discrepancy. ADTRAN has maintained its stability within the peer group by most measures, and EPS is no exception.

Peer Derived Value (PDV)
                                                                                                                     Preemium/Discount
                                                         Historical                                                      of Historical
                                  5-Year Avg            Variation vs.          Frwrd (09-10)       Current Variation     Variation to
                                   Multiple              Wghtd Avg               Multiple           vs. Wghtd Avg     Forward Variation              Price

Companies                      Price to Earnings
ADC Telecom                                  38.16                      1.92              24.76                   1.73                  1.11                  8.65
Adtran                                       21.06                      1.06              16.53                   1.15                  0.92                 22.66
Alcatel                                       17.18                     0.86                                                                                  3.37
BigBand                                       15.10                     0.76               32.14                  2.24                  0.34                  4.04
Ciena                                        21.49                      1.08                                                                                 12.29
Cisco Systems                                20.84                      1.05              14.54                  1.01                  1.03                  21.04
Ericsson                                     15.43                      0.78              12.87                  0.90                  0.86                   9.30
Extreme Networks                             35.18                      1.77               17.69                 1.23                  1.43                   2.57
Corning Inc.                                 18.48                      0.93              12.80                  0.89                  1.04                  15.79
JDS Uniphase                                                                              24.02                  1.68                      -                  5.85
Juniper Networks                             32.57                      1.64              25.58                  1.79                  0.92                  24.47
Motorola                                      31.93                     1.61               14.05                 0.98                  1.64                   7.03
Nokia                                         15.47                     0.78               11.81                 0.82                  0.94                  12.95
Polycomm                                     22.27                      1.12              15.83                   1.10                  1.01                 22.71
Sycamore Networks                              5.71                     0.29              45.66                  3.19                  0.09                   3.08
Tellabs                                       23.12                     1.16              20.54                  1.43                  0.81                   6.59
Simple Average                               15.39                                        16.81                                        0.87
Weighted Average                             19.89                                        14.33
                                                                                          -28%
Figure 16.7 (continued)


Peer Derived Value (PDV)
                                                                                                                 Preemium/Discount
                                                     Historical                                                      of Historical
                               5-Year Avg           Variation vs.          Frwrd (09-10)       Current Variation     Variation to
                                Multiple             Wghtd Avg               Multiple           vs. Wghtd Avg     Forward Variation    Price

Companies                  Relative P/E
ADC Telecom                                 2.26                    2.13               1.35                 1.71               1.24             8.65
Adtran                                      1.28                    1.21               0.93                 1.18               1.02            22.66
Alcatel                                                                                                                                         3.37
BigBand                                     0.86                    0.81               1.70                 2.16               0.37
Ciena                                                                                                                                          12.29
Cisco Systems                                1.21                   1.13               0.83                 1.05               1.08            21.04
Ericsson                                    0.92                    0.86               0.74                 0.93               0.92             9.30
Extreme Networks                            4.46                    4.19               0.92                  1.16              3.60             2.57
Corning Inc.                                 1.15                   1.08               0.74                 0.94               1.15            15.79
JDS Uniphase                                2.50                    2.35                1.37                1.74               1.35             5.85
Juniper Networks                            1.94                    1.82               1.45                 1.84               0.99            24.47
Motorola                                    1.57                    1.47               0.93                  1.18              1.25             7.03
Nokia                                       0.64                    0.60               0.66                 0.84               0.71            12.95
Polycomm                                    1.34                    1.26               0.90                  1.14               1.11           22.71
Sycamore Networks                           1.92                    1.80               2.68                 3.40               0.53             3.08
Tellabs                                     1.39                    1.31                1.17                1.48               0.88             6.59
Simple Average                              1.65                                       0.99                                    1.16
Weighted Average                            1.06                                       0.79
352   •   Relational Valuation: The Industry Matrix Workbook and Peer Derived Value



variation; indeed, at 1.24, ADCT’s relative P/E makes the stock attractive com-
pared with that of its peer group. ADTRAN, fittingly, has maintained its peer
group relationship based on relative P/E better than any rival. ADTRAN’s 1.28
historical relative P/E is 1.21 times the historical peer group–weighted relative
P/E of 1.06; its forward relative P/E of 0.93 is 1.18 times the forward peer group–
weighted relative P/E of 0.79. Altogether, with a 1.02 premium of historical varia-
tion to forward variation, ADTRAN is almost exactly where it has always been
in relation to its peer group.


Finishing the PDV Calculation
Notice that the current price for each stock is repeated in each of the PDV grids.
To determine our PDV value, we need to adjust the price by premium or discount
of historical variation to forward variation for each of the five inputs (P/S, P/BV,
P/CF, P/E, and relative P/E). We will then average the five adjusted prices to
determine our PDV. The process is illustrated in Figure 16.8.
      In the following discussion, remember that the exhibit shows a truncated
version of our normal PDV worksheet; typically these values would be 12 to 20
rows to the right, and the column values would be much further along in the
alphabet. In the upper right of this PDV worksheet snippet, in column A we
repeat our company names. In the next five columns, B through F, for each com-
pany we multiply actual price times our premium/discount of historical variation
to forward variation
      In the actual worksheet as opposed to this truncated snippet, for each com-
pany we multiply actual price times our premium/discount of historical variation
to forward variation for price to sales in column J), price to book value (column
K), price to cash flow (column L), price to earnings (column M), and relative
price to earnings (column N).
      Returning to our exhibit, to calculate Peer Derived Value for each company,
in column H we average the five values in columns B through F. In column I we
show the actual price, and in column J we quantify the relationship between PDV
and actual price.
      Let’s return to our test case companies one last time. For ADCT, which at
the time of this exercise was priced at $8.65, we have calculated a Peer Derived
Value of $12.68. In cell J5, we see that this equates to 1.47, or a 47% premium
to actual price. Based on our relational valuation work, ADC could well be that
company we cited when we jumped off on this exercise: “The ADCT shares,
which typically trade at a premium to the peer group, now trade at a
discount.”
Figure 16.8

Having laid the groundwork of assessing historical to current peer relationships on five metrics, within the PDV process we apply that information and use it to
adjust the current price to uncover peer value discrepancies.

Peer Derived Value (PDV)
                                                                                                                                                 Premium/Dscnt
                                                                                                           Peer Calculated                         to Current
                                                    Mutiplied by Current Price                                  Value           Current Price         Price
Companies                         P/S            P/BV         P/CF         P/E             Rel P/E
ADC Telecom                             14.56         5.59        22.90          9.61           10.76                  12.68              8.65             1.47
ADTRAN                                  21.58        15.45        23.11        20.80            23.20                  20.83             22.66             0.92
Alcatel                                  6.72         2.53         3.35             -               -                    4.20             3.37             1.25
BigBand                                  5.03         5.19         8.14          1.37               -                    4.93             4.04             1.22
Ciena                                   22.83         8.72            -             -               -                    7.89            12.29             0.64
Cisco Systems                           22.80       20.96         25.63        21.74            22.74                  22.77             21.04             1.08
Ericsson                                12.46         9.57         8.97          8.03            8.57                    9.52             9.30             1.02
Extreme Networks                         4.72         2.55         5.85          3.68            9.26                    5.21             2.57             2.03
Corning Inc.                            13.43        18.24        20.23        16.43            18.20                   17.30            15.79              1.10
JDS Uniphase                            18.28         6.34       (8.57)             -            7.90                    4.79             5.85             0.82
Juniper Networks                        25.04        13.87        14.28        22.45            24.20                  19.97             24.47             0.82
Motorola                                 7.90         6.58         2.98         11.51            8.78                    7.55             7.03             1.07
Nokia                                   18.73        14.74        13.26        12.22             9.25                  13.64             12.95             1.05
Polycomm                                24.13        15.85        29.65        23.02            25.16                  23.56             22.71             1.04
Sycamore Networks                        2.00         2.11            -          0.28            1.63                    1.20             3.08             0.39
Tellabs                                  6.91         3.93         6.14          5.34            5.83                    5.63             6.59             0.85
Average
354   •   Relational Valuation: The Industry Matrix Workbook and Peer Derived Value



      ADTRAN, steady as a rock, trades pretty much where it normally trades
within its peer group. We have arrived at a PDV of $20.83, compared with a cur-
rent price of $22.66. This relationship equates to 0.92, or roughly an 8% discount
to current price. While we have not uncovered an attractive value discrepancy
with ADTN, neither have we uncovered an overvaluation situation.


Peer Derived Value of the Communications Technology Group
In our final illustration, Figure 16.9, we show a snippet from the upper left corner
of the PDV worksheet. In it are our conclusions: a summary of the PDV for each
stock in column C and the PDV’s relation to its actual price in column E. At the
bottom we have averaged this relationship for the 16 companies, and the average
is 1.05. Give or take normal statistical variation, that’s fairly close to 1. And that’s
because we’ve created an insular and near-perfect closed loop, right?
      Yes, and Hucklebery Finn’s the Duke and the Dauphin really are related to
French Royalty. Our system is no more a closed loop than anything within the
market, excluding perhaps the market itself. Along the way, we’ve had to throw
out unprofitable or hysterically valued outliers; we begin with pro forma valua-
tions, which are necessarily arbitrary. Our forward component is based on a cas-
cade of estimates, not hard data.
      Nonetheless, when we perform this exercise, we consistently arrive at an
average value that is within 15% of 1.0, either to the upside or downside. We
believe we have fashioned a somewhat closed loop. More broadly, we are confi-
dent that quantifying the historical valuation to forward valuation of a company
within its peer group—in other words, establishing a relational valuation—has
valid informational content to contribute to the valuation process.


Risks and Limitations of the PDV Process
Acknowledging (or positing) that PDV has valid informational content brings us
to our next question: how much should PDV contribute to the calculation of dol-
lar value of the asset and the asset decision? We’ll cut to the chase by saying it
should be something substantial, but something less than 25%.
      While we can calculate a PDV value, we believe we should distinguish
between dollar value largely determined by the company’s own operations and
values, and dollar value rendered purely from a relational situation. With com-
parable historical and DFCF, we are significantly reliant on carefully modeled
data to help calculate value. With PDV, we are relying less on company-specific
                                                                     Peer Derived Value     •   355


Figure 16.9

Peer Derived Value can identify value discrepancies, but it is merely part of the overall valuation
scheme, which also blends historical comparable and discounted free cash flow valuation.

Peer Derived Value (PCV)
                                                                               Premium/Dscnt
Companies                                     PDV             Current Price    to current price
ADC Telecom                                          12.68                8.65               1.47
Adtran                                               20.83               22.66               0.92
Alcatel                                                4.20               3.37               1.25
BigBand                                                4.93               4.04               1.22
Ciena                                                  7.89              12.29               0.64
Cisco Systems                                        22.77               21.04               1.08
Ericsson                                               9.52               9.30               1.02
Extreme Networks                                       5.21               2.57               2.03
Corning Inc.                                          17.30              15.79                1.10
JDS Uniphase                                           4.79               5.85               0.82
Juniper Networks                                     19.97               24.47               0.82
Motorola                                               7.55               7.03               1.07
Nokia                                                13.64               12.95               1.05
Polycomm                                             23.56               22.71               1.04
Sycamore Networks                                      1.20               3.08               0.39
Tellabs                                                5.63               6.59               0.85
Average                                                                                     1.05




data (mainly that included in the Big Five comparables) and relying more on the
relationship of the company to its peer group.
      This heightens risk that we can misinterpret a value trap as an attractive
value discrepancy. A company about to implode can “supernova” its earnings
through reckless pricing, prize asset dispositions, and other forestalling actions
that only intensify the final collapse. In a vacuum, a PDV much higher than the
trading price might signal a huge value opportunity; in context, we might see the
company is actually in a tailspin.
      A related difficulty with relational valuation is simply that companies don’t
stand still in their industries. A company need not be in this supernova phase to
send a false PDV signal; it may simply be struggling in the face of intensifying
competitive headwinds. And a company that seems expensive on a relational
valuation basis may be pulling away from its competition, thus earning more of
a premium status than investors accorded the stock in the past.
      Finally, and perhaps most importantly, the PDV process is based on actual
price adjustment, not price calculation. An efficient market supporter might
356   •   Relational Valuation: The Industry Matrix Workbook and Peer Derived Value



argue that the actual price already reflects all market inputs and that to further
adjust it is a distortion. Our counterargument is that relational valuation is use-
ful; that alternative methods (creating a peer group index) is complex, not vali-
dated, and dependent on unverifiable inputs; and that quantifying the intuitive,
however imperfect, is superior to relying only on intuition.
       Altogether, we believe it is appropriate to treat PDV as one aspect of the
valuation equation. In practice, it represents about 15% to 20% of our valuation
equation. Fairly, that begs the question: when is it 15%, and when is it 20%?
While our answer is a bit arbitrary, it too is intuitive. When a PDV has great
variation in its components and/or when the PDV varies significantly from actual
price, we are at least somewhat suspect of the informational content, and we
assign it a 15% weighting.
       ADC Telecom’s PDV is accorded a 15% weighting in the dollar value of the
asset calculation. Its components send divergent signals ($14.56 from P/S, $5.59
from P/BV), and its PDV is 47% higher than its actual price. ADTRAN, by con-
trast, has remarkably steady PDV; four of the five price-based metric relationships
are in the low $20s, or consistent with its price at the time of this exercise. Alto-
gether, its PDV of $20.83 was within 8% of its actual price at the time of the
exercise. In other words, ADTRAN is trading more or less where it always has in
relation to the peer group.
       If that were to begin to change, we would regard it as a significant event—
much more so than a change in ADC’s intrinsically less stable relationship to the
group. Therefore, we give ADTRAN’s PDV the higher weight of 20% in the dol-
lar-value-of-the-asset process. Should ADTN’s PDV begin skewing meaningfully
up or down, we could consider going as high as 25%, because any significant
change in this rock-steady relationship would be big valuation news indeed.


PDV in Valuation and the Asset Decision
Now that we’ve calculated Peer Derived Value for each asset, we take that infor-
mation—calculated on the industry matrix workbook—and bring it back to the
individual company workbook. At this point, you should be able to do so without
excessive hand holding on our part.
     We now have three individual contributors to our dollar value of the asset
formulation: comparable historicals, discounted free cash flow, and Peer Derived
Value. Our current (as of mid-2009) weighting in the valuation scheme is 55%
DFCF, 30% comparable historicals, and 15% PDV. Shade it as you see fit.
                                                        Peer Derived Value   •   357



      What might influence adjustments in the weighting scheme? When the
market is running hot, it is presumably looking forward; greed is in the driver’s
seat, and investors want to know what firms plan to do tomorrow. We might
shade the DFCF weighting slightly higher.
       When the market is grinding to a halt or retracing, and fear takes over,
investors scan past performance seeking proven earners. In this circumstance, we
might shade the comparable historicals mix higher. In periods in which techno-
logical innovation is shaking the group and normal value relationships are dis-
solving and reforming, we could look a bit harder at PDV.
      In our discussion of the stock value worksheet, we walked through the final
adjustments steps, including determining target price and adjusting forecast per-
formance for the market risk factor (beta). Recall we left a placeholder for PDV
on each stock value worksheet, to be included (along with values for comparable
historicals and DFCF) in the weighting scheme used in determining dollar value
of the asset. Really, though, we’re winding down the formal modeling process.
This page intentionally left blank
           CONCLUSION:
           DOLLAR VALUE
           OF THE ASSET




America is an entrepreneurial cornucopia, and over the decades I have watched
countless companies come into being—and a fair amount fall away. The Internet
represents that rare thing, a genuine secular earthquake that has been rearrang-
ing the business landscape. In a turbulent but nonetheless fruitful period for
American capitalism, new firms have arisen to meet new needs, and technology
has developed from being a discreet sector to a deeply embedded yet evolving
process within each company. Along the way, investor interest in the stock mar-
ket has surged.
      Consistent with the experience across Wall Street, numerous analysts have
jumped in and out of our firm over the years in pursuit of more lucrative oppor-
tunities. First as my firm’s associate research director and later as director of
research, I helped shepherd new analysts into the industry. Along the way, I par-
ticipated in launching many new and existing companies into coverage. Given
the investing excitement attached in particular to the Internet but really much
more widespread, I found myself initiating coverage in nearly every industry and
sector in preparation for or in assistance to industry-specific analysts.
      Over these years, my first step has always been to build a model, survey and
stress-test it, and only then issue an opinion on the underlying equity. In a second
derivative variation on the journalist’s aphorism—“I need to read what I’ve writ-

                                                                                359
360   •   Conclusion: Dollar Value of the Asset



ten in order to know what I’m thinking” —I discovered that I need to model in
detail so I can determine what I’ll write and only then learn what I am
thinking.
      Along the way, we’ve trialed all the different ways to skin a cat: top-down
and bottom-up modeling and valuation; percentage-of-sales and segment-driven
operating income statements; comparable historical and peer group valuation;
dividend discount and discounted free cash flow valuation. Eventually, what
emerged was a modeling and valuation template that was rigorous and consistent
within the company composite but sufficiently flexible to accommodate new data
inputs and new ways to skin those cats.
      That said, this is the model now, and we’re sticking to it. We find it works
best, at least fresh out of the box, when it is constructed according to the template.
Anecdote time: Three days before Christmas one year, I yanked the pieces of my
son’s bike from the box, too cheap to pay Toys“R”Us to assemble it for me. It
looked simple enough, and I didn’t have time to read what the manufacturer
(what does he know?) had to say about putting it together. Two days later, in a
frazzled state, I wheeled the semidilapidated bike under the Christmas tree. As
the front wheel swung to the side into (wobbly) kickstand pose, the headlamp
gave me a doleful look. At that point I realized I would have saved two days of
work if I’d “wasted” a few minutes reading the manual. This is our manual, and
we don’t recommend straying from the script.
      Analysts who’ve gone through any combination of formal and informal
training have been saturated in financial theory, highly useful if at times conflict-
ing. The conflicts in financial academia provide an interesting backdrop. But any
such musings are dispelled by the ringing of the phone and the client wanting to
know, simply: buy it or sell it? If it seems we demand an almost pedantic precision
about the process, it’s because we can’t afford wasted motion if we are going to
incorporate and integrate all our data on all our worksheets and get them talking
among themselves while still giving the analyst time to talk to clients.
      We hope our work over the preceding chapters has helped you answer that
endlessly recurring question at the end of the phone line. Along the way, we’ve
accomplished much: we have modeled the income statement and other financial
statements, and we’ve used historical and modeled data for the company and its
peer group to calculate value of the asset. We have made provisions to accom-
modate the cycle and the ever-widening schism between GAAP and non-GAAP
results. We’ve incorporated and weighted all the market’s valuation mechanisms,
and we’ve done so in a disciplined and flexible manner. Hopefully, and this is
paramount, we’ve left no loose ends in the valuation process.
                                          Conclusion: Dollar Value of the Asset   •   361


      For all that, permit yourself a sigh of satisfaction—but not much more.
Wrapping up college days, I recall dropping off a final paper—the final paper, in
fact—with a teaching assistant. I let loose a theatrical sigh of completion; college
was really done. The TA snickered. “Your parents are allowed that sigh,” he said.
“You haven’t finished anything.” Two months later my sighs were more of the
anguished variety as I scanned the employment pages and fretted landing an
interview.
      Your handful of models is really much like that sheepskin: (well done, by
the way, for both, but . . .) only a start. Numbers assembly is merely the first step
in modeling. Once the basic framework is constructed, the real work—calibra-
tion—begins. Eventually, what you are modeling is less the number itself and
more the degree to which your process derives a value that varies from the num-
ber. Once you determine consistency for that variance, you’re positioned to make
the necessary adjustments. Now multiply this task by the line items in the indi-
vidual model, and multiply again by the number of individual models. The fin-
ished model finds you not at the finish but at the starting line.
      The role of modeling in asset analysis is significant, but it is far from the
only element. When new analysts start at our company Argus, we tell them that
the asset analysis process has four broad buckets: (1) financial statement model-
ing, (2) valuation analysis, (2) company knowledge, and (4) industry knowledge.
Yet even that represents only a few tools in the tool kit. Beyond individual asset
analysis lies the interrelations of assets: the balance of buy, sell, and hold ratings
for the analyst; the asset management process for the portfolio manager. Simul-
taneously, investment professionals are charged with interacting with clients, a
healthy dose of marketing, and maneuvering through the office politics and the
back-office minutiae that never make it onto the job description but somehow eat
up big chunks of the day.
      In Michael Pollan’s excellent The Omnivore’s Dilemma (Penguin, 2007), the
author introduced a lay audience to the concept of the Holon (which Wikipedia
attributes to Arthur Koestler). The Holon is something that is a complete and
integrated system in its own right yet simultaneously a subsystem or component
of a greater whole. Koestler referred to Holons as autonomous, self-reliant parts.
We’ve tried to approach modeling with the goal of creating a self-contained sys-
tem, a “stable form able to withstand disturbances,” but always within the knowl-
edge that a model is an intermediate form contributing to “the proper functionality
for the larger role” —that is, the analyst’s role.
      Throughout this process, we’ve been putting numbers on just about every-
thing. So, here’s the final question: what percentage of the analyst’s job is model-
362   •   Conclusion: Dollar Value of the Asset



ing? For once, we defer. The financial services industry is simply too open ended
for any one answer to suffice. I know a hedge fund trader who plays, not the
bounce on the news, but the next day’s rebound off the bounce; that’s all he
trades. I interviewed a prospective analyst who in his current job started each day
in cash, traded equities all day, and ended in cash; worn out by his daily grind,
he was 31 years old. I also know portfolio managers who change their holdings
much less frequently than Standard & Poor’s changes the constituents in the S&P
500, and others who still make pencil marks on charts.
      Given the changes in the financial service industry in recent years, includ-
ing the increase in high-velocity program trading and quant strategies based on
complex algorithms, the meticulous modeler can feel a bit like Bartleby Scrivener,
dipping his quill in the inkwell while computers whir in the background. The
inkwell set may have felt some malicious glee when the quant “rocket scientists”
drove their collateralized rockets straight into the mountainside—“without let-
ting off the throttle,” as one wistful PM said to me—in the summer and fall of
2008. Such smugness is out of place, as old-fashioned financial managers cannot
point to much better performance in that historically bad period.
      The nature of the game has changed, and a mere 58 percent market decline
is no more likely to dislodge growing reliance on computer-driven trading and
quant strategies than the slide rule is likely to take back the desktop from the
personal computer. For all that, meticulous modeling is not just vital to the mar-
ket; we’d argue that it is secure in the market.
      Most quant strategies have an exhaustive backlog of data but only a wispy
forward element. Dig through the algorithm for that forward element and you’ll
find the consensus—which even now is built on individually modeled expecta-
tions. There is the risk that cost cutting could squeeze the last few humans out of
the process, and that digital trend compilation will replace the necessarily subjec-
tive mix of hard modeling and industry assessment that informs the analysis
process. Should the outlook become purely dependent on machine-generated
trend analysis, then no mountainside may be safe.
      Our outlook is not so dire, if only because the bookends of the industry—
greed and fear—need to find, respectively, confirmation and succor in a human
face. The financial services industry—with trillions of dollars at stake, and even
now with hundreds of thousands of employees—will remain a multifaceted world
with a wealth of styles, approaches, theories, and gimmicks. Even though the
financial data stream is now a binary blur, we think the industry will always find
a place for those with a feel for the numbers.
          BIBLIOGRAPHY




Damodaran, Aswath, Investment Valuation: Tools and Techniques for Determining
   the Value of Any Asset, (New York: John Wiley & Sons, Inc., 1996)
Cottle, Sidney, Roger F. Murray, Frank E. Block, Graham and Dodd’s Security
     Analysis, Fifth Edition (New York: McGraw-Hill Book Company, 1988)
El-Erian, Mohamed, When Markets Collide (New York, McGraw-Hill, 2008)
Horngren, Charles T., Gary L. Sundem, John A. Elliot, Introduction to Financial
    Accounting, Fifth Edition (Englewood Cliffs, N.J.: Prentice-Hall, 1993)
Gwatney, James D., Richard L. Stroup, Economics: Private and Public Choice, Sev-
    enth Edition (Fort Worth: The Dryden Press, Harcourt Brace College Pub-
    lishers, 1995)
Keiso, Donald E., Jerry J. Weygandt, Intermediate Accounting, Eighth Edition
     (New York: John Wiley & Sons, Inc., 1995)
Maginn, John L., Donald L. Tuttle, Managing Investment Portfolios (Boston: War-
    ren, Gorham & Lamont, 1990)




                                                                            363
364   •   Bibliography



Newton, Harry, Newton’s Telecom Dictionary, Eighteenth Edition (New York:
    CMP Books, 2002)
Puzmanova, Rita, Routing and Switching: Time of Convergence? (Boston: Addi-
    son-Wesley Publishing Company, 2002)
Reilly, Frank K., Investment Analysis and Portfolio Management, Fourth Edition
      (Fort Worth: The Dryden Press, Harcourt Brace College Publishers, 1994)
Ross, Stephen A., Randolph W. Westerfield, Bradford J. Jordan, Fundamentals of
     Corporate Finance (Chicago: Irwin, 1991)
Solnik, Bruno, International Investments, Third Edition (Reading, Mass.: Addi-
     son-Wesley Publishing Company, 1996)
White, Gerald I., Ashwinpaul C. Sondhi, Dov Fried, The Analysis and Use of
    Financial Statements (New York: John Wiley & Sons, Inc., 1994)
Yamarone, Richard, The Trader’s Guide to Key Economic Indicators (New York:
    Bloomberg Press, 2004)
              INDEX




A                                      and stockholders’ equity, 226           in operating cash flow, 178
Absolute return, 294–295               (See also Mergers & acquisitions        in stock value worksheet, 264
Accelerated depreciation, 14                 [M&A])                          Amortized inputs, in cash flows, 187
Accounting changes (income           Action alerts, 292, 293                 Analog Devices Inc., 170, 183
     statement), 16                  Action sheet, 285, 295                  Analysis and Use of Financial
Accounting standards, 14–15, 20      Actual contribution, 113                      Statements (Sondhi, White,
  FAS 123R, 14–15, 20, 23, 31, 44,   Actual price adjustment, 355–356              and Fried), 176
         46                          ADC Telecom, 23, 26–27, 70–72, 340      Analysts:
  FAS 142, 23, 30, 159, 226, 345     Add-backs, 54, 59                         buy-side, 144
  IFRS, 14, 104                      Adjusted earnings, 205, 279               compensation of, 284
  and impairments, 28                Adjusted earnings per share, 227          GAAP vs. non-GAAP
Accounting taxes, cash taxes vs.,    Adjusted line items, 28 (See also Pro            conventions of, 28
     168                                  forma (PF) line items)               sell-side, 138, 143–144, 158–159
Accounts payable (A/P), 153–154,     Adjusted results, 20, 226 (See also     Annual balance sheets, quarterly vs.,
     311                                  Pro forma results)                       138
Accounts receivable (A/R), 311       Adjusted segment income, 81             Annual growth, 69
  and inventories, 153–154           ADR (see American Depository            Annual updates:
  and receivables turnover, 155           Receipt)                             and discounted free cash flows,
Acquired intangibles, 40 (See also   ADTRAN Inc., 10–12, 196–200,                     268–269
     Intangibles, amortization of)        202–207, 340                         and historical comparable
Acquisitions:                        American Depository Receipt                      valuation, 209, 213
  EBITDA for, 166                         (ADR), 104, 106–107                A/P (see Accounts payable)
  and goodwill, 23                   Amortization:                           Appreciation, 241, 293
  line item for overpayment, 30        on income statement, 14               A/R (see Accounts receivable)
  and segment presentation,            of intangibles, 20, 40, 44            ARIMA-12 (time series method), 98
         118–119                       in margin model, 40                   ASP (see Average selling price)



                                                                                                             365
366    •   Index



Asset(s):                                  metrics, 311–312                     Capital appreciation percentage, 293
  acquisitions, 118–119                    modeling, 137–139 (See also          Capital asset pricing model
  blended value of, 269, 270, 273                Ratio and valuation                  (CAPM), 240
  buy-rated, 274                                 worksheet)                     Capital spending, 178, 201, 264
  discontinued, 16, 19, 36                 and ordinary least square            Capital-intensive companies, 172
  dollar value of (see Dollar value              regressions, 130–132           Capitalism, 100
          of asset)                     Basic EPS, 19, 36                       Capitalism, Socialism and
  and goodwill, 23                      Basic shares, 36, 55, 56                      Democracy (J. Schumpeter),
  impairment to, 30, 35–36              Benchmark earnings, 332                       100
  price data, 196                       Benchmarks:                             CAPM (capital asset pricing model),
  and price percentage change              industry, 285                              240
          worksheet, 285                   market, 7                            Cap-weighted sequential change,
  price performance, 284, 285              and price/performance analysis,            321
  tangible, 159                                  293                            Cash (balance sheet), 312
  turnover of, 157, 159, 243            Beta, 241                               Cash accounting, 14
Asset decision:                            adjusting risk with, 221             Cash and investments (margin
  and market return, 273                   and forecast asset value, 267              model), 50
  and peer derived value, 356–357       Big Five historical comparables,        Cash (cash conversion) cycle, 311
  and price/performance analysis,             329–330, 333                        concept of, 161
          296                           Blended value of asset, 269, 270, 273     defined, 158
  and risk-adjusted return, 274         Bloomberg (company), 191                  and payables turnover, 160
  and stock value worksheet, 219,       Bloomberg, Michael, 238                   shortening, 312–313
          221                           Bloomberg.com, 230                      Cash flow(s):
Asset valuation, xiii                   Boeing, 85                                amortized inputs in, 187
Asset value:                            Bonds, 229                                annual vs. quarterly, 138
  based on DDM, 230                     Book accounting, 14                       and company comparison
  based on peer group                   Book taxes, 16                                   worksheet, 317
          relationships, 331–332 (See   Book to bill, 92                          discreet, 254–256, 261, 268
          also Peer derived value       Book value:                               forecasting, 232
          (PDV))                           of assets, 313                         future, 260
  forecast, 194, 267                       and company comparison                 growth rate for, 238
  and goodwill, 227                              worksheet, 317                   modeling, xii–xiii
  investor interest in, 70                 of debt, 239                         Cash flow from investing (CFI), 178
  terminal, 256                            of equity, 313                       Cash flow from operations (CFO),
Assets under management (AUM),             market value vs., 238                      177
      325                                  and return on total capital, 169     Cash flow ratios, 174, 176–182
Asset-tracking software, 292, 293          (See also Price/book value)          Cash flow statements:
AUM (assets under management),          Book-and-build companies, 92              data for OLS and NE worksheet
      325                               Book-and-ship companies, 92                      from, 126
Automotive Week (trade journal),        Bottom-up analysis, xii                   direct vs. indirect, 176
      84–85                             BRAC nations, 217                         modeling, 137–139 (See also
Average premium to blended stock        BRIC nations, 217                                Ratio and valuation
      value, 304                        Broadcom, 194, 195                               worksheet)
Average return, 285                     Buffett, Warren, 232                    Cash flow-based valuation, 176–177,
Average selling price (ASP), 91, 95     Bulge bracket firms, 86                       188
Aviation Week (trade journal), 85       Business cycles, 121, 132               Cash on hand, 46
                                        Buy-rated assets, 274                   Cash payments, one-time, 265 (See
B                                       Buy-sell triggers, 292–293                    also Nonrecurring items)
Balance sheet:                          Buy-side analysts, 144                  Cash taxes:
  annual vs. quarterly, 138                                                       book taxes vs., 16
  data for OLS and NE worksheet         C                                         and ROIC, 168
        from, 125                       CAGR (see Compound annual               Celestica, 146–151
  income statement vs., 49                  growth rate)                        CFA Institute, 153
  and margin model, 44, 49, 50          CapEx, 266                              CFI (cash flow from investing), 178
                                                                                                 Index     •   367



CFP (cash flow from operations),         U.S.-based, 49                          in discounted free cash flow, 225,
      177                                valuation of, 283                              255
Changes in accounting treatment,       Company comparison worksheet,             and valuation, 229
      16                                    284, 303–320                       Continuing operation lines (see Pro
Chemical Week (trade journal), 84        and balance sheet metrics,                  forma (PF) line items)
Ciena Corp., 57–58, 66–68, 88–90               311–312                         Convertible debt, 29, 36, 57, 61
Cisco, 20–22, 46–48, 235–237             data for, 306                         Corning, 16–18, 52, 174, 181, 244–
COGS (see Cost of goods sold)            earnings in, 306–310                        249, 251–252, 325
Collusion, 28                            and forecast values, 304–305          Corporate performance, 229
Commercial paper (CP), 155               and margins, 312–316                  Corporate risk-taking, 227
Common stock equivalents (CSEs),         purpose of, 303                       Corporate strategy, 88, 117,
      19, 57                             return measures on, 313, 317                173–174
Companies:                               revenues in, 310–311                  Corporate tax rates, 16, 35
   book-and-build, 92                    and valuations, 317–320               Cost against revenue, 54
   book-and-ship, 92                   Company equity workbooks (see           Cost items adjustments, 98–100
   capital-intensive, 172                   Individual equity workbooks)       Cost of debt, 239, 241
   comparisons of, 280 (See also       Company knowledge, 172                  Cost of equity:
          Company comparison           Company presentations, 87, 91             calculating, 240–241
          worksheet)                   Company strategy, 88, 117, 173–174        and discounted free cash flow,
   consensus-gathering, 28             Company-only data, 296                           267
   and debt/capitalization, 174        Comparables, 191 (See also                weighted, 241
   distressed, 256                          Historical comparables)            Cost of goods sold (COGS):
   energy, 86                          Comparisons:                              data on, 70
   engineering-intensive, 46             company, 280 (See also Company          on income statement, 13–14
   foreign (see Foreign companies)             comparison worksheet)             in margin model, 39, 43
   globally diverse, 16, 301             historical, 81 (See also Historical     non-GAAP, 29
   gross margins of, 164 (See also             comparable valuation)             and percentage of difference, 80
          Gross margin)                  peer group, xiii–xiv, 174               stock option compensation in,
   growth, 241, 311                      sequential, 13                                 31
   with high fixed costs, 100, 172     Compensation:                           Costs:
   with high variable costs, 100         of analysts, 284                        fixed, 101
   immature, 310                         stock, 55–56                            impairment, 81–82
   as information sources, 51, 60,       stock option, 14, 29, 31–34, 36–        on income statements, 49
          74, 85–91, 98                        37, 44, 46                        interest, 15, 49–51
   information-industry, 55            Competition, 226                          manufacturing, 80
   joint-venture (see Joint-venture    Compilation columns (margin               net interest income, 50, 51
          (JV) companies)                   model), 60–62                        operating (see Operating costs)
   large, 15–16, 23, 66                Component-based DFCF, 260–267             restructuring, 81–82
   low-debt, 56                        Compound annual growth rate               stock option, 46
   in mature industries, 256                (CAGR), 313                          variable, 101
   in nascent or dynamic industries,     and losses, 307                       Coverage ratio (times interest
          256                            and normalized earnings,                    earned), 173
   partner, 16                                 122–124                         CP (commercial paper), 155
   public, 9, 31, 64–66                  and OLS growth rate, 126, 128,        Creative destruction, 121
   regulated, 241                              130                             Credit rating agencies, 172
   reporting by, 9–10, 31, 63–66,        unadjusted, 122–123, 310              CSE interest add-back, 59
          74                           Conference calls, 85–87, 91, 117        CSEs (see Common stock
   resources-intensive, 69             Consensus earnings estimates, 20              equivalents)
   return on assets for, 169           Consensus managers, 28                  Currency translations:
   seasonal, 69 (See also Seasonal     Consensus-gathering companies, 28         exchange rates for, 105–106, 112,
          effects)                     Consolidation (joint-venture                     213
   small, 14–15, 66, 325                    companies), 16                       and historical comparable
   start-up, 14, 20, 23                Constant growth rate (g):                        valuation, 213–214
   technology, 46, 69, 86, 313           concept of, 181, 182, 246–247           in industry matrix models, 279
368   •   Index



  for joint-venture companies, 111– DDA (depreciation, depletion, and           and current prices, 304
         112, 116                          amortization), 14                    discount rates in, 238
  and valuations, 104–107            DDM (dividend discount model),             and dollar value of assets, 188
Current accounts, 154                      229–231                              and forecast growth, 242–247
Current earnings, 344                Debt:                                      and growth periods, 247–249
Current liabilities, 154                book value of, 239                      historical comparables vs., 138–
Current prices:                         convertible, 29, 36, 57, 61                    139, 223–224
  in comparables valuation, 196         cost of, 239, 241                       limitations of, xiii, 224–229
  and discounted free cash flow,        junk, 239                               per share, 243–247
         304                            in margin model, 49, 50                 in present value modeling,
  in price/earnings ratio, 193–194      market value of, 239                           232–233
Current ratio, 154, 312                 as percentage of total capital, 240     in ratio valuation worksheet,
Current variation, 341, 342             speculative, 239                               138–139
Customer base, 158                      and WACC, 241–242                       ratios for, 176–181
Cyclical stocks, 141                    weighted cost of, 241                   sensitivity of, 267–268
                                     Debt/capitalization (debt/cap) ratio,      in stock value worksheet, 219,
D                                          229                                         269–275
D&A (depreciation and                   concept of, 174, 175                    three-stage, 257–260
      amortization), 266                and stock compensation, 56              and trade working capital, 153
Damodaran, Aswath, 177, 231          Debt/equity ratio, 173                     two-stage, 251–257
Data:                                Depreciated inputs, 187                    types of, 231–232
  annual, 142                        Depreciation, 13–14, 201, 264              underlying concepts of, 235–238
  for company comparison                and cash flow, 176                      (See also Present value modeling)
         worksheet, 306                 of goodwill, 23, 226                  Discounted present value, 266
  company-only, 296                     in operating cash flow, 178           Discounted terminal value, 268
  cost of goods sold, 70                and ROIC, 168                         Discreet cash flows, 254–256, 261,
  historical (see Historical data)   Depreciation, depletion, and                   268
  income, 70–72                            amortization (DDA), 14             Disruptive innovation, 100
  industry, 95, 98, 283              Depreciation and amortization            Distressed companies, 256
  in industry matrix, 278–279 (See         (D&A), 266                         Dividend(s), 230–231
         also Industry data          DFCF (see Discounted free cash             historical patterns for, 231
         compilations)                     flow)                                in income statements, 37
  market cap, 294                    Diluted EPS, 3, 19, 36, 57–59            Dividend discount model (DDM),
  for OLS and NE worksheet, 125,     Diluted shares, 36, 55, 56, 264                229–231
         126                         Direct cash flow statements, 176         Dividend payments, 105, 108
  operating income, 66               Disaggregation:                          Dividend payout ratio:
  orphaned, 296                         in joint-venture companies, 116         concept of, 180
  price, 196                            of price per unit from number of        retained earnings vs., 180–181
  public, 278–279                             units produced, 95              Dividend yield, 273
  for ratio valuation worksheet, 143    of return on equity, 244              Dollar sign ($) symbol (in Excel),
  regional units, 95                    and unit trends, 91                         126, 145, 307
  revenue, 63–67, 98                 Disclosure:                              Dollar value of asset, 359–362
  segment, 63–67, 74, 83, 117–120       Regulation Fair Disclosure, 63,         contributors to, 356
  stock value worksheet, 220                  86–87                             determining, 188
  on trends, 311                        of segment data, 63–67                  and DFCF, 188, 235
  trends in, 88                      Discontinued operations, 54, 59            and enterprise value to EBITDA,
  unit, 87, 95                       Discount rates, 238–243                           209
  for valuation methodologies, 140 Discount to equity, 231–232                  estimating, 270
  valuations of, 317, 320            Discount to the firm, 231–232, 255         hard, 141
  from Web sites, 238                Discounted cash flow valuation, 177        and HCV, 187–188
  from workbench, 83                 Discounted free cash flow (DFCF),          and historic comparables,
  (See also Information sources)           235–249                                     208–213
Day sales outstanding (DSO), 158–       annual updates, 268–269                 and PEG/PEGY, 206
      159, 311–313                      component-based, 260–267                and price/book value, 198
                                                                                                 Index    •   369



  and price/cash flow, 199–202        EBIT (earnings before interest and      Excel, 6, 42, 108, 268
  and price/free cash flow, 201–203         taxes):                             dollar sign ($) symbol in, 126,
  and price/sales ratio, 194,            data for modeling, 88                          145, 307
        196–197                          in times interest earned, 173          downloads for, 191
  and relative P/E, 203–206           EBITDA margin, 165–166                    F2 (Function 2) key, 268–269
  in stock value worksheet, 219,      EBITDA per share, 189                     and industry matrix, 303–304
        251                           EBITDA/enterprise value (EV),             linking cells in, 144–145
Down cycles:                                166–167, 189                        managing links in, 163–164
  of 2007-2009, 224                   EBT (earnings before taxes), 74         Exchange rates, 105–106, 112, 213
  and current ratio, 153, 312         Economic cycles, 100–101                      (See also Currency
  and goodwill impairment, 227           and cash cycle, 161                        translations)
  market collapse, 86, 227               and historical comparables,
  and price/free cash flow, 201                138–139                        F
  recessions, 124, 136                   and operating leverage, 172–173      FAS (see Financial accounting
  and working capital, 153               and ordinary least square                   standards)
  and working capital/sales ratio,             regressions, 123               FASB (see Financial Accounting
        155                              and payables turnover, 160                  Standards Board)
  (See also Economic cycles)             (See also Down cycles)               FCF (free cash flow), 178–180
DSP (see Day sales outstanding)       Economic Value Added (EVA), 168         FIFO (first in, first out), 14
DuPont ROE, 170–171, 243–246          Economies, 16 (See also United          Financial accounting standards
                                            States economy)                          (FAS):
E                                     EDGAR (electronic data gathering,          FAS 123R (stock option
Earnings:                                   analysis, and retrieval), 7, 9              compensation), 14–15, 20,
  adjusted, 23, 28, 205, 279          EE (see Equity income)                            23, 31, 44, 46
  benchmark, 332                      EMP (Ericsson Mobile Platforms),           FAS 142 (impairments), 23, 30,
  in company comparison                     116                                         159, 226, 345
        worksheet, 306–310            Energy companies, 86                    Financial Accounting Standards
  current vs. future, 344             Engineering-intensive companies,               Board (FASB), 16, 23, 28,
  distrust of, 167                          46                                       226–227
  estimates of, 20                    Enterprise value (EV) (see EBITDA/      Financial services industry, 362
  GAAP, 23, 28, 192                         enterprise value)                 Financial statement accounting, 14
  growth in, 307, 310                 EPS (see Earnings per share)            Financial statements:
  inputs for, 3                       Equity:                                    of public companies, 9
  myth of, 140–141                       cost of, 240–241, 267                   and ratio and valuation
  non-GAAP, 313                          as percentage of assets, 243–244               worksheet, 143–151
  normalized (see Normalized             as percentage of total capital,         (See also specific types of
        earnings)                              239–240                                  statements)
  pro forma, 23, 28, 347                 of stockholders (see Stockholders’   First in, first out (FIFO), 14
  revenues vs., 187, 194                       equity)                        Fisher, Irving, 177
  and seasonal effects, 99            Equity income (EE):                     Five-year projections, 247
  segment, 74                            and joint-venture companies,         Fixed asset turnover, 159
Earnings before taxes (EBT), 74                116                            Fixed costs, 101
Earnings per share (EPS):                as line item, 16                     Fixed-income securities, 239
  adjusted, 227                          in margin model, 54, 55              Flextronics International, 161
  basic, 19, 36                          workbench modeling of, 109,          Forecasting, 84–85
  diluted, 3, 19, 36, 57–59                    111                               asset value, 194, 196, 267
  GAAP, 56–59, 180, 227               Equity turnover, 159                       cash flows, 232
  growth, 122–123, 249                Equity value, 304                          DFCF per share, 243–247
  in margin model, 55–59              Ericsson, 103–110, 117, 118,               growth, 242–247
  measures of, 307                          213–217                              for unit shipments, 112
  pro forma, 36–37, 57–59, 136        Ericsson Mobile Platforms (EMP),           values, 304–305
  projections of, 307                       116                               Foreign companies:
Earnings-based valuation, 176–177,    Estimates, earnings, 20                    ADRs of, 213–217
     188                              Euro, 104, 111–112, 116, 213               comparison of, 301
370    •   Index



   and corporate tax rates, 16          General Electric (GE), 64–65, 117   Growth rates:
   and cost line items, 49              Genuine Parts Co. (GPC), 231          constant (see Constant growth
   modeling for, 103–110                Global business, 279                        rate)
   P&L for, 105–116                     Global growth, 312                    and discounted free cash flow,
Foreign currency (see Currency          Globally diverse companies, 16,             257–262
      translations)                          301 (See also Foreign            economic, 256
Foreign pretax income, 35                    companies)                       EPS, 249
Form 20-F (SEC), 105                    Goodwill, 23                          historical, 67, 69, 122, 132
Forward P/Es (price/earnings            Goodwill depreciation, 226            OLS (see Ordinary least squares
      multiples), 190–191, 193          Goodwill impairment, 23, 30                 growth rates)
Forward variation, 341–344, 349,          consequences of, 23–28              terminal, 256
      352                                 and equity turnover, 159            (See also Compound annual
Free cash flow (FCF), 178–180             line item for, 30                         growth rate [CAGR])
Fried, Dov, 152, 176                      and stockholders’ equity, 172,    Growth stocks, 198, 256
Future cash flows, 260                           226–228
Future earnings, 344                      tax effects of, 35                H
                                        GPC (Genuine Parts Co.), 231        Hard stops, 292
G                                       Gross domestic product (GDP),       Hedge funds, 143, 226
g (see Constant growth rate)                 247, 256                       Historical asset price data, 191–192,
G&A costs (see General &                Gross margin:                            196
       administrative costs)              concept of, 164                   Historical cash flow:
GAAP, IFRS vs., 104                       GAAP vs. PF, 46                     and operating cash flow, 77
GAAP line items:                          inputs for, 70                      ROIC for, 167
   basic EPS, 56                          pro forma, 46                     Historical comparable valuation,
   diluted EPS, 56–59                     as profit driver, 88                   187–217
   earnings, 23, 28, 192                  and seasonal effects, 44            adjustments for, 190–191
   earnings per share, 56–59, 180,        in workbench, 98                    for ADRs of foreign companies,
          227                           Gross profit, 29                             213–217
   gross margin, 46                     Growth:                               and annual updates, 209, 213
   interest income, 52                    annual, 69                          dollar value of assets in,
   main operating costs, 46               cash flow, 238                             193–213
   net income, 36, 40, 180                and discounted free cash flow,      for equity value, 304
   operating costs, 30–32, 40, 44                242–249                      importance of P/Es in, 187–189
   operating income, 32, 35, 40, 46,      earnings, 307, 310                  PEG/PEGY ratios in, 206–208
          128                             economic, 256                       preparing grid for, 191–193
   P&L operating income, 81–82            EPS, 122–123, 249                   price/book value in, 198–200
   pretax income, 40, 49, 51–52, 128,     global, 312                         price/cash flow in, 199, 201
          130                             historical, 67–69, 122, 132         price/earnings ratio in, 192–193
   and pro forma line items, 28–29        and maturity, 247                   price/free cash flow in, 201–203
   return measures, 313                   metrics for, 242–247                price/sales ratio in, 194, 196–197
   and segment operating income,          modeling, 69                        relative P/E in, 203–206
          81–82                           normalized, 261, 264                time frames for, 189–190
   and stock option compensation,         obsession with, 231–232           Historical comparables, 191, 219
          31–34                           periods of, 247–249                 for ADRs of foreign companies,
   taxes, 35–36, 40, 52, 54               phases of, 257, 258, 260                   213–217
GAAP results, 37, 226, 279                quarterly, 69                       balanced approach to, 141–142
   investor acceptance of, 28             revenue, 92, 310                    Big Five, 329–330, 333
   pro forma results vs., 20–23           sales, 92                           discounted free cash flow vs.,
GDP (see Gross domestic product)          sequential, 69, 92                         138–141, 232–233
GE (see General Electric)                 share, 266                          and limitations of P/Es, 140–141
General & administrative (G&A)            (See also Growth rates)             and overreliance on P/Es,
       costs, 43, 44 (See also Selling, Growth assumptions, 122                      139–140
       general, & administrative        Growth companies, 241, 311            present value modeling vs.,
       (SG&A) costs)                    Growth metrics, 242–247                      223–224
                                                                                                Index     •   371



  (See also Book value; Cash flow;     Income:                                   (See also Margin model; Segment
         Price/book value; Price/         equity, 16, 54, 55, 109, 111, 116            revenue modeling)
         cash flow; Price/earnings;       foreign pretax, 35                  Income statement presentation, xvi,
         Price/sales; Relative P/E;       interest, 14, 49–52, 74                   2, 42, 103
         Sales)                           modeling, xii–xiii                  Income tax expense, 52
Historical comparisons, 81                net (see Net income)                Income taxes, 15–16
Historical data:                          operating (see Operating income)    Incremental margin, 43–44
  for company comparison                  pretax (see Pretax income)          Indentures, 229
         worksheet, 306                   segment, 73, 74, 81, 82 (See also   Indexes, 7
  in income statement modeling,                 Segment operating income)     Indirect cash flow statements, 176
         28–37                         Income data, 70–72                     Individual company workbook, 118
  for joint-venture companies, 111     Income statement(s), 5–37              Individual data worksheet, 296
  ordinary least square regressions       accounting conventions for, 28      Individual equity workbooks, 5–6,
         for, 123–124                     annual updates of, 209, 213,              277–278
  presentation of, 67                           268–269                       Industrialized nations, 16
  price, 191–192, 295                     balance sheet vs., 49               Industry analysis, 283
  for segment revenue modeling,           below-the-line line items in, 16–   Industry analysis firms, 86
         67                                     19, 36                        Industry benchmarks, 285
  and segment revenue                     costs on, 49                        Industry data, 95, 98, 283, 296, 301
         presentation, 119                data for OLS and NE worksheet       Industry data compilations:
  sources of, 7, 9                              from, 125                        in industry matrix workbook,
  and stock option compensation,          dividends in, 37                             296–301
         31–32                            and GAAP, 20–37                        and price/performance analysis,
Historical dividend pattern, 231          and goodwill, 23–28, 30, 35–36               284
Historical growth rates, 67–69, 122,      layout of, 220                         purpose of, 280
     132                                  modeling (see Income statement      Industry knowledge, 172
Historical model, 42–44                         modeling)                     Industry matrix workbook,
Historical norm, 159                      and operating income, 32, 35              277–281
Historical P/E (price/earnings            and pretax income, 32, 35              company comparisons in, 280
     multiple), 190–193                   standard line items of, 13–19                (See also Company
Historical percentage growth, 67,         and stock option compensation,               comparison worksheet)
     69                                         31–32                            data in, 278–279
Historical segment revenue, 67            structure of, 225                      and equity turnover, 159
Historical variation:                     and taxes, 35–36                       and Excel, 303–304
  and forward variation, 349, 352         (See also Normalized earnings;         and fixed asset turnover, 159
  premium discount of, to forward               Percentage-of-difference         and individual company
         variation, 341–344                     modeling; Workbench)                   workbook, 118
  weighted average vs., 334–340        Income statement modeling, xvi,           industry data compilations in,
Historical years, 40, 42–44                  5–37                                      280, 284, 296–301
Hold rating, 295                          changing status quo for, 9–10          industry data in, 296
Holon, 361                                and conventions in reporting,          margin ratios in, 163
                                                28                               margins in, 313
I                                         historical data amendments in,         organization of, 279–281
Ibbotson, 193                                   28–37                            peer derived value in, 280 (See
IFRS (see International Financial         historical data sources for, 7, 9            also Peer derived value
      Recordings Standards)               nonrecurring and noncash items               worksheet)
Immature companies, 310                         in, 19–28                        price performance grid for, 280
Impairment:                               operating and pretax income                  (See also Price and
   to assets, 30, 35–36                         amendments in, 32, 35                  performance analysis)
   costs of, 81–82                        pro forma line items in, 19–28         purpose of, 278–279
   and GAAP taxes, 52                     and standard line items, 13–19         quarterly trends worksheet in,
   to goodwill (see Goodwill              stock option adjustments in,                 321, 324
         impairment)                            31–34                            query page for, 280, 284
   of stockholders’ equity, 345           structuring the model, 10–13           and segment reporting, 119–120
372    •   Index



   segment worksheet in, 280,           Investors:                               operating costs in, 44, 46–48
          320–322                          conference calls with, 85–87, 91,     operating income in, 44, 46–48
   weightings worksheet in, 280,                  117                            per-share earnings in, 55–59
          321, 325–330                     momentum, 107, 293                    pretax income in, 51–53
Industry standards, 312–313             Investor’s Business Daily, 292           taxes in, 52, 54
Inflated stock, 227                     IR (investor relations), 85              year-over-year comparison in,
Inflation, 14                                                                           59–60
Information management, 296             J                                      Margin percentages, 132
Information sources:                    JDSU (company), 40, 41, 227            Margin ratios, 163–167
   companies, 51, 60, 74, 85–91, 98     Joint-venture (JV) companies, 16       Margins:
   conference call transcripts, 91         currency translations for, 111–       and company comparison
   formal company presentations,                 112, 116                               worksheet, 312–316
          91                               disaggregation in, 116                EBITDA, 165–166
   industrywide, 95, 98                    and equity income, 116                gross (see Gross margin)
   Internet, 84–85, 87                     and margin model, 54–55               incremental, 43–44
   for revenue modeling, 64                pro forma line items for, 29          net, 165, 243
   for segment data, 63–67                 workbench modeling for, 109,          operating, 70, 80, 164–165
   trade journals, 84–85, 95, 98                 111–116                         pretax, 165
   (See also Data; Financial            Juniper Networks, 44, 45, 87, 117,       segment, 66, 73
          statements; Reporting)              320                              Market benchmarks, 7, 221
Information-industry companies, 55      Junk debt, 239                         Market capitalization, 56, 228, 294
Innovation, disruptive, 100             JV companies (see Joint-venture        Market capitalization weighting,
Insider information, 86                       companies)                            294, 321, 325–330
Intangibles, amortization of, 20, 40,                                          Market collapse, 86, 227
      44                                K                                      Market perceptions, 187
Interest costs:                         Kroner, 104, 112–113, 116, 213–214     Market proxy, 269
   in margin model, 49–51                                                      Market risk premium, 240–241,
   and operating income, 15             L                                           267
Interest coverage, 173                  Large companies, 15, 16, 23, 66        Market risk premium assumption,
Interest income:                        Last in, first out (LIFO), 14               267
   GAAP, 52                             Liabilities, current, 154              Market share, 296, 301
   in margin model, 49–51               LIFO (last in, first out), 14          Market value:
   and operating income, 15             Liquidity, 152–156, 311                  book value vs., 238
   and percentage of difference, 74     LM Ericsson Telefon (see Ericsson)       of debt, 239
   pro forma, 52                        Losses:                                  of equity, 239
Internal liquidity ratios, 152–156         and actual contribution, 113        Market-weighting function, 321
International Financial Reporting          and compound annual growth          Master limited partnerships
      Standards (IFRS), 14, 104                   rate, 307                         (MLPs), 231
Internet:                                  proportional, 16                    Maturation of companies:
   as information source, 84–85, 87,       and risk taking, 292                  and change in reporting
          105                           Low-debt companies, 56                          segments, 117
   query page linked to, 284                                                     and slower growth, 247
Intuition, 356                          M                                      Mergers & acquisitions (M&A):
Inventories, 311                        M&A (see Mergers & acquisitions)         EBITDA for, 166
Inventory collection period, 158        Main operating costs, 46                 and stockholders’ equity, 226
Inventory days outstanding, 158         Majority partners, 16, 55                (See also Acquisitions)
Inventory turnover, 158                 Manufacturing costs, 80                Minority interest (MI), 16, 54, 109,
Investment Analysis (White, Sondhi,     Margin model, 39–62                         111 (See also Joint-venture
      and Fried), 152                     compilation columns in, 60–62             (JV) companies)
Investment decisions, 28 (See also        forward periods in, 40, 42           MLPs (master limited partnerships),
      Asset decision)                     interest cost in, 49–51                   231
Investment Valuation (Aswath              interest income in, 49–51            Modeling, xiii–xv
      Damodaran), 231                     margin analysis for, 39–41             of growth, 69
Investor psychology, 292                  modeled quarters in, 42–45             for income statements, 1–2, 5–7,
Investor relations (IR), 85               net income in, 54–55                          9–10, 65
                                                                                             Index     •   373



  of operating costs, 100–101             on income statements, 23, 28–32,  Operating leverage:
  percentage-of-revenue, 10, 74                  42                            concept of, 172–173
  refinement process for, 92              in margin model, 44                  in workbench, 100–101
  role of, 361                          NOPAT (net operating profit after   Operating margin, 70, 80, 164–165
  segment-based, 64                          taxes), 168                    Operating metrics, 20 (See also
  and valuation, 1                      Normalized earnings (NE),                 GAAP results; Pro forma
  in workbooks, 5–7                          124–136                              results)
  (See also specific models by            concept of, 124–125, 132          Operating performance, dividend
        name)                             method of, 133–136                      discount model vs., 230
Momentum investors, 107, 293              and OLS growth rates, 125–132     Operating profit:
Motorola, 7, 8, 74–81, 87, 92, 95–97,     and OLS regressions, 132–136        pro forma, 32–33, 35
     126–132                              worksheet concept for, 132          segment, 73–74
MSN Money, 7, 8                         Normalized growth, 261, 264         Operating progress, 230–232
Multiples of revenue, 166               Nortel, 227                         Operating strategy, 101
                                        NYSE (New York Stock Exchange),     Operations:
N                                            104                              discontinued, 54, 59
Nasdaq Exchange, 104                                                          impression of improvement in,
NE (see Normalized earnings)          O                                              56
Net income:                           OLS growth rates (see Ordinary        Operations and maintenance
  defined, 36                               least squares growth rates)           (O&M), 178
  and FAS 123R, 15                    OLS regressions (see Ordinary least Ordinary least squares (OLS)
  GAAP, 36, 40, 54–55, 180                  squares regressions)                  growth rates:
  on income statement, 130            O&M (operations and                     balance sheet data, 130–132
  in margin model, 54–55                    maintenance), 178                 and compound annual growth
  in normalized earnings, 136         The Omnivore’s Dilemma (M.                     rate, 126, 128, 130
  in operating cash flow, 178               Pollan), 361                      income statement data, 128–130
  pro forma, 29, 36, 40, 54–55        One-time items, 20, 42, 265 (See also   worksheet setup and method,
  in stock value worksheet, 264             Nonrecurring items)                      125–128
Net income per share, 19 (See also    1/1 ratio (trading), 106              Ordinary least squares (OLS)
      Earnings per share (EPS))       Opening price, 292                          regressions:
Net income taxes, 16                  Operating cash flow, 177                in company comparison
Net interest cost (income), 50        Operating costs, 80                            worksheet, 310
Net interest income (cost), 51           categories of, 43                    concept of, 123–124
Net margin, 165, 243                     GAAP, 30–32, 40, 44, 46              and normalized earnings,
Net operating profit after taxes         as line item, 14, 29–32                     132–136
      (NOPAT), 168                       main, 46                             in stock value 2 worksheet, 264
New business originations, 225           in margin model, 43, 44, 46–48     Orphaned data, 296
New York Stock Exchange (NYSE),          modeling of, 100–101               Outlier performances, 325
      104                                pro forma, 29–32, 44, 46           Overvaluation of stock, 342, 344,
Noncash items:                           and stock option compensation,           345
  defined, 20                                  31–32
  exclusion of, 20, 23, 81               (See also specific types of costs) P
  inclusion of, 81–82                 Operating earnings to GAAP            Parent-company identity, 109
  on income statements, 20, 23,             percentage, 227–228             Partner companies, 16
         28–32                        Operating efficiency ratios, 155,     Paste-special (Excel command), 108
  in margin model, 44                       157–162                         Payables collection period, 160
  and stockholders’ equity, 229       Operating income, 46                  Payables turnover, 160
Noncontrolling interest (see             data on, 66                        Payment terms, 13
      Minority interest (MI))            GAAP, 32, 35, 40, 46, 81–82, 128 P/BV (see Price/book value)
Non-GAAP line items, 28, 313 (See        as line item, 15–16, 30–32, 35     P/CF (see Price/cash flow)
      also Pro forma (PF) line items)    in margin model, 44, 46–48         PDV (see Peer derived value)
Nonmajority partners, 16                 and percentage of difference, 74   PDV worksheet (see Peer derived
Nonoperating items, 49, 74               pro forma, 30–32, 35, 46, 81             value worksheet)
Nonrecurring items:                      profit & loss, 74, 77, 81–82       P/E multiple (see Price/earnings
  and goodwill impairment, 226           segment, 73–82                           multiple)
374   •   Index



Peer derived value (PDV), 283          Per-share price, 191                       market cap-weighted return in,
   discrepancies in,